shades of green - Utrecht University Repository - Universiteit Utrecht

Loading...
SHADES OF GREEN Spatial and temporal variability of potentials, costs and environmental impacts of bioenergy production

SHADES OF GREEN Spatial and temporal variability of potentials, costs and environmental impacts of bioenergy production. Floor van der Hilst, May 2012

The research reported in this thesis was carried out at the Science, Technology and Society Group of Utrecht University and at the Valorisation of Plant Production Chains group of Wageningen University and Research Centre. Financial support was granted by the Dutch Government and Shell Research Foundation for the fourth Mitigation of Emissions project (ME4) within the Climate changes Spatial Planning programme (Dutch: Klimaat voor Ruimte). In addition, the research has been partly funded by the Biorenewable Resources Platform, SASOL and GEF UNEP/FAO/UNIDO. ISBN: Print: Cover design: Foto: Copyright:

978-90-8672-054-5 BOXpress Thijs van Himbergen, Floor van der Hilst (image: Athos Boncompagn) Frank Moll © 2012, Floor van der Hilst

SHADES OF GREEN Spatial and temporal variability of potentials, costs and environmental impacts of bioenergy production SCHAKERINGEN IN GROEN Ruimtelijke en temporale variabiliteit in de potentiëlen, kosten en milieu-impacts van bioenergie productie (met een samenvatting in het Nederlands)

NUANCES DE VERDE Avaliacao espacial e temporal das potencialidades, custos e impactos ambientais da producao de Bioenergia (contem resumo e conclusoes em portugues)

ВІДТІНКИ ЗЕЛЕНОГО Просторово-часова зміна потенціалів, вартість та екологічний вплив виробництва біоенергії (з резюме українською мовою)

Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof. dr. G.J. van der Zwaan, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op woensdag 16 mei 2012 des ochtends te 10.30 uur door Floortje van der Hilst geboren op 28 september 1981 te Utrecht

Promotoren:

Prof. dr. A.P.C. Faaij Prof. dr. J.P.M. Sanders

Table of contents LIST OF FIGURES

IX

LIST OF TABLES

XIV

UNITS AND ABBREVIATIONS

XVII

1

INTRODUCTION 1.1 1.2 1.3 1.4

21

CURRENT AND FUTURE ENERGY SUPPLY THE ROLE OF BIOENERGY NEED FOR REGIONAL SPECIFIC ASSESSMENTS AIM AND THESIS OUTLINE

2 POTENTIAL, SPATIAL DISTRIBUTION AND ECONOMIC PERFORMANCE OF REGIONAL BIOMASS CHAINS; THE NORTH OF THE NETHERLANDS AS EXAMPLE

22 22 25 27 33

2.1 INTRODUCTION 35 2.2 CASE STUDY DESCRIPTION 36 2.2.1 Study region 36 2.2.2 Biomass potential in the region 37 2.2.3 Bioenergy chains 38 2.3 METHOD 40 2.3.1 NPV calculations for crop production 41 2.3.2 Cost of ethanol 43 2.3.3 NPV and costs of feedstock differentiated for soil suitability 44 2.4 RESULTS 46 2.4.1 Cost of biomass 51 2.4.2 Costs of ethanol 52 2.4.3 Sensitivity analysis 55 2.5 DISCUSSION 57 2.5.1 Method and input data 57 2.5.2 Results 58 2.6 CONCLUSIONS 59 2.7 ACKNOWLEDGEMENTS 61 2.8 APPENDIX: INPUT DATA FOR CALCULATION OF ECONOMIC PERFORMANCE OF BIOENERGY CROPS62 2.8.1 Cultivation 62 2.8.2 Transport 69 2.8.3 Conversion 69 2.8.4 Regional economic factors 70 V

3 SPATIAL VARIATION OF ENVIRONMENTAL IMPACTS OF REGIONAL BIOMASS CHAINS 3.1 INTRODUCTION 3.2 METHODS 3.2.1 GHG emissions 3.2.2 Soil quality 3.2.3 Water use and water quality 3.2.4 Biodiversity 3.2.5 Overview methods and integration of results 3.3 RESULTS 3.3.1 GHG emissions 3.3.2 Soil erosion 3.3.3 Water use and water quality 3.3.4 Biodiversity 3.3.5 Integrated results 3.4 DISCUSSION AND CONCLUSIONS 3.5 ACKNOWLEDGEMENTS 3.6 APPENDIX I: METHODS AND INPUT DATA 3.6.1 Parameters for calculation of the GHG emissions 3.6.2 Parameters for calculation of the impact on soil quality 3.6.3 Parameters to calculate the impact on water use and water quality 3.6.4 Parameters to calculate the impact on biodiversity 3.7 APPENDIX II: ADDITIONAL RESULTS 3.7.1 Additional results of GHG emissions 3.7.2 Additional results of impact on soil quality 3.7.3 Additional results of impact on water quality and quantity 3.7.4 Additional results for impact on biodiversity

73 75 76 78 79 82 84 86 87 87 89 90 92 93 95 98 99 99 102 107 108 113 113 114 115 117

4 SPATIOTEMPORAL LAND USE MODELLING TO ASSESS LAND AVAILABILITY FOR ENERGY CROPS – ILLUSTRATED FOR MOZAMBIQUE

121

4.1 INTRODUCTION 4.2 METHODOLOGY 4.2.1 Drivers of land use change 4.2.2 Scenario approach 4.2.3 Land use modelling 4.3 RESULTS 4.3.1 Developments in demand, productivity and land requirements 4.3.2 Deterministic spatial modelling results 4.3.3 Uncertainty in spatial modelling results

123 125 125 126 128 132 132 135 140

VI

DISCUSSION AND CONCLUSION 4.4 4.5 ACKNOWLEDGEMENTS 4.6 APPENDIX: INPUT DATA AND MODEL RULES CASE STUDY MOZAMBIQUE 4.6.1 Land use change drivers in Mozambique 4.6.2 Land use modelling Mozambique 5 SPATIOTEMPORAL COST-SUPPLY CURVES FOR BIOENERGY PRODUCTION IN MOZAMBIQUE 5.1 INTRODUCTION 5.2 METHODOLOGY AND DATA INPUT 5.2.1 Land availability 5.2.2 Feedstock production costs 5.2.3 Conversion costs 5.2.4 Transportation costs 5.3 RESULTS 5.3.1 Feedstock production costs 5.3.2 Conversion costs 5.3.3 Transportation costs 5.3.4 Total supply cost to international market 5.4 DISCUSSION 5.5 CONCLUSIONS 5.6 ACKNOWLEDGEMENTS 5.7 APPENDIX: ADDITIONAL RESULTS PRIMARY TRANSPORT COSTS 6 INTEGRATED SPATIOTEMPORAL ANALYSIS OF AGRICULTURAL LAND USE, BIOENERGY PRODUCTION POTENTIALS AND RELATED GHG BALANCES IN UKRAINE 6.1 INTRODUCTION 6.2 METHODS 6.2.1 Future demand 6.2.2 Scenarios 6.2.3 Land use change model 6.2.4 GHG emissions 6.3 RESULTS 6.3.1 Total agricultural land balance 6.3.2 Spatially explicit results 6.3.3 GHG emissions 6.4 DISCUSSION 6.4.1 PLUC model 6.4.2 GHG calculations VII

144 147 149 149 153 163 165 166 166 167 171 173 178 178 184 184 186 194 197 198 199 203 205 206 207 207 210 215 221 221 223 227 233 233 234

Data requirements 6.4.3 6.5 CONCLUSIONS 6.6 APPENDIX: INPUT DATA GHG EMISSIONS AND ADDITIONAL RESULTS 6.6.1 Input data 6.6.2 Additional results 7

SUMMARY AND CONCLUSIONS 7.1 7.2 7.3 7.4 7.5 7.6

237 238 240 240 242 245

RESEARCH CONTEXT AIM AND RESEARCH QUESTIONS SUMMARY OF THE RESULTS MAIN FINDINGS AND CONCLUSIONS RECOMMENDATIONS FOR FURTHER RESEARCH MARKET AND POLICY RECOMMENDATIONS

246 247 249 253 263 264

SUMÁRIO E CONCLUSÕES

267

РЕЗЮМЕ ТА ВИСНОВКИ

291

SAMENVATTING EN CONCLUSIES

317

REFERENCES

341

DANKWOORD

365

CURRICULUM VITAE

369

VIII

List of Figures FIGURE 1.1: COMPLEX INTERACTIONS OF BIOENERGY WITH THE SOCIO-ECONOMIC AND ENVIRONMENTAL CONTEXT AT MICRO AND MACRO SCALE, TAKEN FROM IPCC (2011). FIGURE 2.1: INDIVIDUAL CONTRIBUTIONS OF COST ITEMS AND BENEFITS TO NET PRESENT VALUE (NPV) OF INDIVIDUAL CROPS, CROP ROTATIONS, AND PERENNIAL ENERGY CROPS EXCLUDING SUBSIDIES. FIGURE 2.2: NET PRESENT VALUE OF PERENNIALS, TYPICAL ROTATION SCHEMES AND ROTATIONS SCHEMES INCLUDING AN EXTRA SHARE OF SUGAR BEET (ES) FOR DIFFERENT SOIL SUITABILITY CLASSES (EXCLUDING SUBSIDIES). FIGURE 2.3: MAP OF ΔNPV (= NPV OF CURRENT LAND USE - NPV PERENNIAL ENERGY CROPS) FOR THE WHOLE AGRICULTURAL AREA OF THE NORTH OF THE NETHERLANDS. NEGATIVE VALUE (GREEN AREAS) INDICATES WHERE MISCANTHUS HAS A HIGHER NPV THAN CURRENT LAND USE. ALL COST ITEMS ARE INCLUDED AND SUBSIDIES ARE OMITTED. FIGURE 2.4: COST SUPPLY CURVES FOR VARIOUS CROPS IN THE NORTH OF THE NETHERLANDS FOR THE TOTAL OF AGRICULTURAL LAND IN THE REGION. THE FIRST ‘STEP’ IN THE CURVES INDICATES THE COST OF BIOMASS PRODUCED ON VERY SUITABLE SOILS, THE SECOND FOR HIGH SUITABLE… THE LAST STEP OF EACH CURVE INDICATES THE COST OF BIOMASS PRODUCED ON VERY MARGINALLY SOILS (LEFT). COST SUPPLY CURVE OF MISCANTHUS BASED ON LAND AVAILABILITY FROM ΔNPV (NET PRESENT VALUE) AND DISTRIBUTION OVER SOIL SUITABILITY AND THE POTENTIAL RELATED TO THE LAND AVAILABILITY ACCORDING TO THE REFUEL STUDY (RIGHT). FIGURE 2.5: SPATIAL DISTRIBUTION OF SUGAR BEET PRODUCTION COSTS (TOP) AND MISCANTHUS PRODUCTION COSTS IN €/GJ. FIGURE 2.6: COST OF ETHANOL PRODUCTION FROM VARIOUS FEEDSTOCK IN THE NORTH OF THE NETHERLANDS COMPARED TO PETROL PRICES FOR VARIOUS OIL PRICE LEVELS (US$/BARREL). LEAST COST FEEDSTOCK PRODUCED ON VERY SUITABLE SOILS ARE INCORPORATED. (CAPEX: CAPITAL EXPENDITURES, O&M: OPERATIONS AND MAINTENANCE COSTS). FIGURE 2.7: SENSITIVITY ANALYSIS FOR NET PRESENT VALUE (NPV), COST OF BIOMASS, AND COST OF ETHANOL OF MISCANTHUS AND SUGAR BEET. KEY PARAMETERS, DISCOUNT RATE, ENERGY PRICES, LABOUR WAGES, YIELD LEVELS, COMMODITY PRICES AND EFFICIENCY OF CONVERSION, ARE VARIED BETWEEN 50% AND 200% OF THE ORIGINAL VALUE. FIGURE 3.1: OVERVIEW OF THE BIOETHANOL CHAIN AND THE ENVIRONMENTAL IMPACTS ASSESSED, THE INDICATORS USED AND THE MODELS APPLIED. THE BOTTOM BOXES INDICATE THE TYPE OF RESULT THAT IS GENERATED FROM THE ASSESSMENTS. FIGURE 3.2: GHG EMISSIONS DUE TO SOC CHANGES AND DURING LIFECYCLE OF BIOETHANOL PRODUCTION FROM MISCANTHUS AND SUGAR BEET. FOR COMPARISON, THE AVERAGE GHG EMISSIONS OF PETROL OVER THE LIFECYCLE IS DEPICTED. FIGURE 3.3: Δ SOC WHEN CURRENT LAND USE IS CONVERTED TO MISCANTHUS (LEFT) OR TO SUGAR BEET (RIGHT). FIGURE 3.4: Δ GHG WHEN CURRENT LAND USE IS CONVERTED TO MISCANTHUS (LEFT) OR TO SUGAR BEET (RIGHT). FIGURE 3.5: Δ WIND EROSION RISK IN KG SOIL/HA/Y WHEN CURRENT LAND USE IS CONVERTED TO MISCANTHUS (LEFT) OR TO SUGAR BEET (RIGHT).

IX

25 47

48

49

51 54

55

56

86

87 88 89 90

FIGURE 3.6: Δ WATER DEPLETION DURING SUMMER (MM) WHEN CURRENT LAND USE IS CONVERTED TO MISCANTHUS (LEFT) OR SUGAR BEET (RIGHT). FIGURE 3.7: Δ NO3 CONCENTRATION (MG/L) WHEN CURRENT LAND IS CONVERTED TO MISCANTHUS (LEFT) OR TO SUGAR BEET (RIGHT). FIGURE 3.8: RISK OF BIODIVERSITY LOSS WHEN CURRENT LAND USE IS CONVERTED TO MISCANTHUS (LEFT) OR TO SUGAR BEET (RIGHT) BASED ON HNV CLUSTER INDICATOR. FIGURE 3.9: INTEGRATED RESULT OF ENVIRONMENTAL IMPACT OF A SHIFT FROM CURRENT LAND USE TO MISCANTHUS (LEFT) OR SUGAR BEET (RIGHT) BASED ON STANDARDISED MAPS OF INDIVIDUAL ENVIRONMENTAL IMPACTS. ST ND FIGURE 3.10: ENERGY REQUIREMENTS FOR 1 AND 2 GENERATION BIOETHANOL PRODUCTION IN THE NORTH OF THE NETHERLANDS. FIGURE 3.11: Δ N2O EMISSIONS WHEN CURRENT LAND USE IS CONVERTED TO MISCANTHUS (LEFT) OR TO SUGAR BEET (RIGHT). FIGURE 3.12: EROSION RISK FOR INDIVIDUAL CROPS ON SANDY SOILS (LEFT) AND CLAY SOILS (RIGHT) OVER THE YEAR. FIGURE 3.13: RISK ON EROSION IN KG SOIL/HA/Y CURRENT LAND USE. FIGURE 3.14: EFFECTIVE PRECIPITATION AND CROP SPECIFIC EVAPOTRANSPIRATION DEVELOPMENT OVER THE YEAR FOR THE CLIMATE CHARACTERISTICS OF EELDE (WEATHER STATION IN GRONINGEN). FIGURE 3.15: WATER DEPLETION (MM) DURING SUMMER (APRIL-SEPTEMBER) FOR CURRENT LAND USE. FIGURE 3.16: WUE FOR MISCANTHUS AND SUGAR BEET BIOMASS (LEFT) AND BIOETHANOL (RIGHT). FIGURE 3.17: Δ N BALANCE IN SOIL WHEN CURRENT LAND USE IS CONVERTED TO MISCANTHUS (LEFT) OR TO SUGAR BEET (RIGHT). FIGURE 3.18: Δ P BALANCE IN SOIL WHEN CURRENT LAND USE IS CONVERTED TO MISCANTHUS (LEFT) OR TO SUGAR BEET (RIGHT). FIGURE 3.19: CHANGE IN MSA VALUE FOR LUC TO MISCANTHUS (LEFT) AND TO SUGAR BEET (RIGHT) OUTSIDE HNV AREAS. FIGURE 4.1: OVERVIEW OF THE MODELLING OF LAND AVAILABILITY FOR BIOENERGY CROPS. FIGURE 4.2: TOTAL FOOD AND NON-FOOD CROP DEMAND IN TIMEFRAME 2005-2030 CONSIDERING THE DEVELOPMENTS IN POPULATION GROWTH, DIETARY INTAKE AND SSR RATIOS. THE ERROR BARS INDICATE THE RANGE IN DEMAND GIVEN THE LOWER AND HIGHER PROJECTIONS FOR POPULATION GROWTH (32 MILLION 36 MILLION PEOPLE IN 2030) (UNDP 2008) AND DIETARY INTAKE (2050 - 2980 KCAL/CAP/DAY IN 2030) (FAO 2003B). FIGURE 4.3: DEVELOPMENT IN CROP AND LIVESTOCK PRODUCTIVITY IN THE BAU AND PROGRESSIVE SCENARIOS IN THE TIME FRAME 2005-2030, NORMALISED FOR THE PRODUCTIVITY LEVELS OF 2005 (2005=1). THE

91 92 93

94 113 114 115 115 116 117 117 118 118 118 131

133

BANDWIDTHS REPRESENT THE RANGE OF THE UNIFORM DISTRIBUTION OF THE STOCHASTIC INPUT OF YIELD DEVELOPMENTS FOR THE BAU AND PROGRESSIVE SCENARIOS. FIGURE 4.4: LAND REQUIREMENTS FOR LIVESTOCK GRAZING AND CROP PRODUCTION FOR THE TIMEFRAME 20052030 FOR THE BAU (LEFT) AND PROGRESSIVE (RIGHT) SCENARIOS, GIVEN THE SAME DISTRIBUTION OVER PRODUCTIVITY CLASSES OF PASTURE AND ARABLE LAND AS IN 2005. THE ERROR BARS REPRESENT THE RANGE IN TOTAL LAND REQUIREMENT GIVEN THE UNCERTAINTIES IN TOTAL DEMAND (FIGURE 4.2) AND PRODUCTIVITY (FIGURE 4.3). FIGURE 4.5 : LAND USE DYNAMICS UP TO 2030 FOR THE BAU (UPPER MAPS) AND PROGRESSIVE SCENARIOS (BOTTOM MAPS).

X

133

135 138

FIGURE 4.6 : LAND AVAILABILITY FOR BIOENERGY CROPS IN 2005, 2015 AND 2030 FOR BAU AND PROGRESSIVE SCENARIOS (BASED ON DETERMINISTIC RUNS). RED AREAS INDICATE THE AREAS THAT ARE NOT AVAILABLE, WHEREAS THE GREEN AREAS ARE THE AREAS AVAILABLE FOR BIOENERGY CROP PRODUCTION. FIGURE 4.7 THE DEVELOPMENT OF LAND AVAILABILITY FOR BIOENERGY CROP PRODUCTION OVER TIME FOR THE BAU (LOWER TREND LINE) AND PROGRESSIVE SCENARIOS (UPPER TREND LINE). FIGURE 4.8: THE DEVELOPMENT OF LAND AVAILABILITY OVER TIME DIFFERENTIATED FOR SUITABILITY CLASSES FOR THE BAU (LEFT) AND THE PROGRESSIVE SCENARIOS (RIGHT). FIGURE 4.9: UNCERTAINTY IN LAND AVAILABILITY FOR BIOENERGY CROPS IN THE BAU SCENARIO FOR THE TIME FRAME 2005-2030. FIGURE 4.10: PROBABILITY OF LAND AVAILABILITY FOR BIOENERGY CROPS AT GRID CELL LEVEL FOR SEVERAL TIME STEPS FOR THE BAU SCENARIO BASED ON STOCHASTIC INPUT VARIABLES OF SEVERAL PARAMETERS (SEE TABLE 4.8). FIGURE 5.1: LEFT SIDE: THE REQUIRED BIOMASS GATHERING AREA GIVEN THE INPUT REQUIREMENTS OF THE CONVERSION PLANT, THE DISTRIBUTION OF AVAILABLE LAND, THE PRODUCTIVITY OF THE LAND AND THE MANAGEMENT FACTOR. THE AVERAGE TRANSPORT DISTANCE IS 2/3 OF THE RADIUS OF THE REQUIRED GATHERING AREA. RIGHT SIDE: THE LEAST COST DISTANCE FROM ANY GRID CELL TO A HARBOUR GIVEN THE COST SURFACE DETERMINED BY THE AVAILABILITY AND THE QUALITY OF ROAD INFRASTRUCTURE. FIGURE 5.2: ROAD INFRASTRUCTURE MOZAMBIQUE, DERIVED FROM ANE (2010). FIGURE 5.3: COST BREAK DOWN OF EUCALYPTUS CULTIVATION FOR AGRO-ECOLOGICAL SUITABILITY LEVELS IN 2010 AND 2030. CULTIVATION COSTS DECREASES BY HIGHER SUITABILITY LEVELS AS SEVERAL COST ITEMS ARE PER HECTARE AND DECREASE PER TON OF PRODUCT WITH HIGHER YIELDS. AS THE MAXIMUM YIELD LEVEL IS HIGHER IN 2030, THE CULTIVATION COSTS PER TON PRODUCT ARE LOWER COMPARED TO 2010 FOR THE SAME SUITABILITY LEVEL. AS SEVERAL KEY COST ITEMS (FERTILIZER APPLICATION, HARVEST AND EXTRACTION) ARE RELATED TO YIELD (AND NOT TO HECTARES) THE COST DECREASE IS ASYMPTOTIC (AND NOT LINEAR). FIGURE 5.4: COST BREAK DOWN OF SUGARCANE CULTIVATION FOR SEVERAL YIELD LEVELS. CULTIVATION COSTS DECREASES BY HIGHER YIELD LEVELS AS SEVERAL COST ITEMS ARE PER HECTARE. AS THE MAXIMUM YIELD LEVEL IS HIGHER IN 2030, THE CULTIVATION COSTS PER TON PRODUCT ARE LOWER COMPARED TO 2010 FOR THE SAME SUITABILITY LEVEL. AS SEVERAL KEY COST ITEMS (FERTILIZER APPLICATION, HARVEST AND LOADING) ARE RELATED TO YIELD (AND NOT TO HECTARES) THE COST DECREASE IS ASYMPTOTIC (AND NOT LINEAR). FIGURE 5.5: SENSITIVITY ANALYSIS OF FEEDSTOCK CULTIVATION COSTS. THE SENSITIVITY OF THE TOTAL COST OF EUCALYPTUS CULTIVATION (LEFT) IS ASSESSED FOR A SUITABILITY OF 75% IN 2010 WHICH CORRESPONDS WITH AS YIELD LEVEL OF 17 ODT/HA/YR. THE SENSITIVITY OF THE TOTAL COST OF SUGARCANE CULTIVATION (LEFT) IS ASSESSED FOR A SUITABILITY OF 75% IN 2010 WHICH CORRESPONDS WITH AS YIELD LEVEL OF 105 TC/HA/YR. FIGURE 5.6: SPATIAL DISTRIBUTION OF COSTS OF EUCALYPTUS FEEDSTOCK PRODUCTION (€/ODT) FOR THE BUSINESS AS USUAL SCENARIO (UPPER THREE MAPS) AND THE PROGRESSIVE SCENARIO (BOTTOM THREE MAPS). FIGURE 5.7: SPATIAL DISTRIBUTION OF COSTS OF SUGARCANE FEEDSTOCK PRODUCTION (€/TC) FOR THE BUSINESS AS USUAL SCENARIO (UPPER THREE MAPS) AND THE PROGRESSIVE SCENARIO (BOTTOM THREE MAPS). FIGURE 5.8: COST OF CONVERSION FROM SUGARCANE TO ETHANOL AND COST FOR PRE TREATMENT OF EUCALYPTUS TO WOOD PELLETS AND TORREFIED PELLETS IN 2010 AND 2030 (EXCLUDING FEEDSTOCK

XI

139 140 140 141

143

174 176

179

179

180

182 183

COSTS).

THE

TOTAL ETHANOL PRODUCTION COSTS ARE REDUCED DUE TO THE REVENUES OF ELECTRICITY

PRODUCTION. FIGURE 5.9: RELATIONSHIP BETWEEN LAND AVAILABILITY, YIELD AND TRANSPORT COSTS FOR EUCALYPTUS (LEFT) AND SUGARCANE (RIGHT) FOR THE PROGRESSIVE SCENARIO IN 2010. FIGURE 5.10: SPATIAL VARIABILITY IN TRANSPORTATION COSTS OF THE END PRODUCTS, (TORREFIED) PELLETS AND ETHANOL, IN €/TON IN 2010, 2020 AND 2030 FOR THE BUSINESS AS USUAL SCENARIO (UPPER THREE MAPS) AND THE PROGRESSIVE SCENARIO (BOTTOM THREE MAPS). FIGURE 5.11: TOTAL COST OF BIOENERGY SUPPLY CHAINS OF (TORREFIED) PELLETS (LEFT) AND ETHANOL (RIGHT) FOR 2010, 2020 AND 2030 FOR THE BUSINESS AS USUAL AND THE PROGRESSIVE SCENARIO, CONSIDERING THE AVERAGE OF ALL AVAILABLE LAND (‘AVERAGE’) AND THE 20% MOST SUITABLE AREAS (‘SUITABLE’). FIGURE 5.12: SENSITIVITY OF TOTAL SUPPLY COST OF TORREFIED PELLETS (LEFT) AND ETHANOL (RIGHT) IN 2030 IN THE PROGRESSIVE SCENARIO FOR CHANGES IN COST OF SUPPLY CHAIN STAGES. THE SENSITIVITY OF THE TOTAL SUPPLY COST FOR THE 20% MOST SUITABLE AREA IS CALCULATED. FIGURE 5.13: TOTAL COST OF SUPPLY CHAIN OF EUCALYPTUS WOOD PELLETS (2010) AND TORREFIED PELLETS (2020 AND 2030) FOR THE BUSINESS AS USUAL AND THE PROGRESSIVE SCENARIO. TOTAL COSTS INCLUDE COST OF FEEDSTOCK CULTIVATION, PRIMARY TRANSPORT, PRE TREATMENT, TRANSPORTATION FORM PLANT TO HARBOUR, STORAGE AND INTERNATIONAL SHIPPING. FIGURE 5.14: TOTAL COST OF SUPPLY CHAIN OF SUGARCANE ETHANOL IN 2010, 2020 AND 2030 FOR THE BUSINESS AS USUAL AND THE PROGRESSIVE SCENARIO. TOTAL COSTS INCLUDE COST OF FEEDSTOCK CULTIVATION, PRIMARY TRANSPORT, CONVERSION, TRANSPORTATION FORM PLANT TO HARBOUR, STORAGE AND INTERNATIONAL SHIPPING. FIGURE 5.15: COST SUPPLY CURVES OF (TORREFIED) PELLETS (LEFT) AND SUGARCANE ETHANOL (RIGHT) FOR 2010, 2020 AND 2030 IN THE BUSINESS AS USUAL AND THE PROGRESSIVE SCENARIO. THE COST SUPPLY CURVE IS A RANKING OF THE SUPPLY POTENTIAL ACCORDING TO THE TOTAL SUPPLY COSTS. FIGURE 5.16: SPATIAL VARIATION IN COSTS OF PRIMARY TRANSPORT OF EUCALYPTUS (€/TON) FOR THE BUSINESS AS USUAL SCENARIO (UPPER THREE MAPS) AND THE PROGRESSIVE SCENARIO (BOTTOM THREE MAPS). FIGURE 5.17: SPATIAL VARIATION IN COSTS OF PRIMARY TRANSPORT OF SUGARCANE (€/TC) FOR THE BUSINESS AS USUAL SCENARIO (UPPER THREE MAPS) AND THE PROGRESSIVE SCENARIO (BOTTOM THREE MAPS). FIGURE 6.1: SCHEMATIC OVERVIEW OF CARBON EXCHANGE BETWEEN ATMOSPHERE, BIOMASS AND SOIL. FIGURE 6.2: SCHEMATIC OVERVIEW OF N PATHWAYS MODELLED IN THIS STUDY. ADAPTED FROM VELTHOF ET AL (2009). FIGURE 6.3: DEVELOPMENT IN DEMAND FOR DOMESTIC PRODUCED FOOD AND FEED IN THE BAU AND PROGRESSIVE SCENARIO IN MILLION TON DRY WEIGHT PRODUCT. FIGURE 6.4: DEVELOPMENTS IN CROP AND PASTURE YIELD AND LIVESTOCK PRODUCTIVITY FOR BAU AND PROGRESSIVE SCENARIO COMPARED TO 2010 LEVELS (2010= 1). FIGURE 6.5: DEVELOPMENTS IN LAND REQUIREMENTS FOR CROP PRODUCTION AND GRAZING FOR BAU AND PROGRESSIVE SCENARIO BASED ON THE FOOD AND FEED REQUIREMENTS, THE YIELD AND EFFICIENCY DEVELOPMENT OF THE TWO SCENARIOS AND ASSUMING THE CURRENT AVERAGE AGRO-ECOLOGICAL SUITABILITY OF ARABLE LAND AND PASTURES. FIGURE 6.6: LAND USE FOR 2010-2020-2030 FOR THE BAU (LEFT) AND PROGRESSIVE SCENARIO (RIGHT). FIGURE 6.7: DEVELOPMENTS IN LAND REQUIREMENTS FOR CROP PRODUCTION AND GRAZING FOR THE BAU AND

184 185

187

189

190

191

192

193 200 201 216 219 222 223

223 226

PROGRESSIVE SCENARIO BASED ON THE SPATIALLY EXPLICIT MODELLING OF THE LAND USE CHANGE IN THE PERIOD 2010-2030 USING THE PLUC MODEL.

227

XII

FIGURE 6.8: DEVELOPMENT IN POTENTIAL ANNUAL BIOMASS PRODUCTION (WHOLE CROP FOR SWITCHGRASS AND GRAIN ONLY FOR WHEAT) FOR THE BAU AND PROGRESSIVE SCENARIO IN PJ/YR. FIGURE 6.9: AVERAGE ANNUAL GHG BALANCE OF TIMEFRAME 2010-2030 COMPARED TO LEVELS OF 2009 FOR THE PROGRESSIVE SCENARIO WHEN ABANDONED AGRICULTURAL LAND IS USED FOR RE-GROWTH NATURAL VEGETATION (TOP), WHEAT FOR BIOETHANOL (MIDDLE), OR SWITCHGRASS FOR BIOETHANOL (BOTTOM) IN 2 TON CO2-EQ/KM PER YEAR. FIGURE 6.10: GRAPHS OF DEVELOPMENT OF CUMULATIVE GHG EMISSIONS FOR THE TIMEFRAME 2010-2030 FOR THE BAU SCENARIO (1), THE PROGRESSIVE SCENARIO (2) AND THE PROGRESSIVE SCENARIO WITH ABATEMENT MEASURES (3). THE THREE VARIANTS CONSIDERED ARE THE USE OF ABANDONED LAND FOR REGROWTH OF NATURAL VEGETATION (A), THE USE FOR WHEAT FOR BIOETHANOL (B) AND THE USE FOR SWITCHGRASS FOR BIOETHANOL (C). FIGURE 6.11: SNAPSHOTS OF ANNUAL N2O EMISSIONS (LEFT), CO2 ABATEMENT (MIDDLE), AND CO2 EMISSIONS (RIGHT) FOR ALL AGRICULTURAL LAND USE AND CULTIVATION OF SWITCHGRASS FOR ETHANOL ON ABANDONED AGRICULTURAL LAND FOR THE YEARS 2010, 2020 AND 2030 FOR THE PROGRESSIVE SCENARIO.

XIII

227

232

233

243

List of Tables TABLE 1.1: OVERVIEW OF THE TOPICS OF THE THESIS CHAPTERS AND THE RESEARCH QUESTIONS ADDRESSED IN THEM. TABLE 2.1: TWO TYPICAL ROTATION SCHEMES FOR SANDY SOILS AND TWO TYPICAL ROTATION SCHEMES FOR CLAY SOILS FOR NORTHERN REGION OF THE NETHERLANDS DERIVED FROM (LEI CBS 2007; VAN DER VOORT ET AL. 2008) EXPRESSED AS PROPORTION OF INDIVIDUAL CROP IN EACH OF THE ROTATIONS. TABLE 2.2: PROPORTION OF LAND THAT COULD BECOME AVAILABLE FOR BIOMASS PRODUCTION IN NORTH OF THE NETHERLANDS ACCORDING TO THREE REFUEL SCENARIOS. TABLE 2.3: CROPS INCLUDED IN THE HELP SYSTEM (HER-EVALUATIE VAN LANDINRICHTINGSPLANNEN – REEVALUATION OF SPATIAL PLANNING) AND NEW CROPS INTRODUCED INCLUDING THEIR RELATIVE SENSITIVITY TO DROUGHT AND WATER DAMAGE. TABLE 2.4: CLASSIFICATION SOIL SUITABILITY AS FUNCTION OF YIELD REDUCTION DUE TO WATER AND DROUGHT STRESS. TABLE 2.5: THE PROPORTION OF LAND THAT IS MORE PROFITABLE UNDER MISCANTHUS OR MORE PROFITABLE UNDER THE CURRENT LAND USES OF ARABLE CROP ROTATIONS ON CLAYEY SOILS AND SANDY SOILS, MAIZE AND GRASS. TABLE 2.6: SHARE OF AREA WHERE MISCANTHUS HAS HIGHER NET PRESENT VALUE THAN CURRENT LAND USE (ΔNPV IS NEGATIVE) IN TOTAL AND FOR DIFFERENT SUITABILITY CLASSES. TABLE 2.7: FIELD OPERATION FOR ANNUAL CROPS BASED ON DE WOLF AND VAN DER KLOOSTER (2006) AND SCHREUDER ET AL. (2008). TABLE 2.8: FIELD OPERATIONS FOR PERENNIAL CROPS. TABLE 2.9: INPUTS, YIELDS AND PRICES OF ANNUAL CROPS. SOURCE: (DE WOLF AND VAN DER KLOOSTER 2006) TABLE 2.10: INPUTS, YIELDS AND PRICES OF PERENNIAL CROPS. TABLE 2.11: PARAMETERS OF FEEDSTOCK TRANSPORTATION USED IN THIS STUDY. TABLE 2.12: DATA ABOUT ETHANOL PLANTS USED IN THIS STUDY (ELSAYED ET AL. 2003; HAMELINCK 2004; HAMELINCK ET AL. 2005A; HAMELINCK AND HOOGWIJK 2007). TABLE 2.13: REGIONAL COST PARAMETERS. TABLE 3.1: PARAMETERS OF THE WEQ EQUATION. TABLE 3.2: ENERGY REQUIREMENTS AND EMISSIONS RELATED TO INPUTS AND FIELD ACTIVITIES DURING CROP CULTIVATION. TABLE 3.3: VALUES USED TO CALCULATE THE CHANGE IN CARBON STOCKS WHEN LAND IS CONVERTED FOR GROWING ENERGY CROPS. VALUES ARE BASED ON THE IPCC (2006) UNLESS OTHERWISE INDICATED. TABLE 3.4: ERODIBILITY INDEXES FOR SOIL CLASSES IN THE NETHERLANDS. A TABLE 3.5: CLIMATE PARAMETERS FOR CALCULATION OF MONTHLY AND ANNUAL CLIMATE FACTOR. TABLE 3.6: MONTHLY VEGETATION FACTORS, SOIL ROUGHNESS FACTORS AND CROP TOLERANCE FOR BLOWING SOIL. TABLE 3.7: CROP EVAPOTRANSPIRATION COEFFICIENT OF CROPS BASED ON (KNMI 2002). TABLE 3.8: REGIONAL AVERAGE CROP EVAPOTRANSPIRATION (BASED ON FIGURES OF TABLE 3.7) AND YIELD OF MISCANTHUS AND SUGAR BEET. TABLE 3.9: VALUES USED TO CALCULATE THE CHANGE IN NITROGEN RELATED EMISSIONS AND PHOSPHOROUS BALANCE WHEN LAND IS CONVERTED FOR GROWING ENERGY CROPS.

XIV

29

37 38

45 45

50 50 63 64 66 67 69 69 71 81 100 101 104 105 106 107 108 108

TABLE 3.10: HNV TYPES AND RELATED CLUSTER INDICATORS BASED ON (ELBERSEN AND VAN EUPEN 2008) TABLE 3.11: THE EXPECTED IMPACT OF BIOENERGY CROPS ON SEVERAL HNV CLUSTER INDICATORS. CHARACTERISTICS OF HNV ARE BASED ON ELBERSEN AND VAN EUPEN (2008). RISK OF BIODIVERSITY LOSS CAUSED BY A SHIFT TOWARDS BIOENERGY CROPS IS BASED ON SEVERAL BIODIVERSITY STUDIES. TABLE 3.12: RELEVANT MSA VALUES DERIVED FROM VAN ROOIJ (2008) AND DORNBURG ET AL. (2008). TABLE 4.1: CURRENT STATUS AND DEVELOPMENT OF KEY DRIVERS OF LAND USE CHANGE FOR THE BUSINESS-ASUSUAL AND PROGRESSIVE SCENARIO. TABLE 4.2: ANNUAL YIELD INCREASE IN % PER ANNUM FOR MAIN CROPS AND YIELD LEVELS IN TON/HA OF TYPICAL CROP TYPES IN TIME FRAME 2005-2015 AND 2015-2030 FOR BUSINESS-AS-USUAL AND PROGRESSIVE SCENARIOS. TABLE 4.3: LIVESTOCK PRODUCTION SYSTEM CHARACTERISTICS. BASED ON FIGURES PROVIDED IN (FAO 2003B; BOUWMAN ET AL. 2005; SMEETS ET AL. 2007). P = PASTORAL SYSTEM M = MIXED SYSTEM. CONTRIBUTION OF GRASS, RESIDUES, SCAVENGING AND FEED CROPS TO TOTAL FEED INTAKE ARE EXPRESSED AS A FRACTION OF TOTAL FEED INTAKE (=1). TABLE 4.4: DEVELOPMENTS IN WOOD DEMAND AND HARVESTING FOR TWO SCENARIOS UP TO 2030. TABLE 4.5: AGGREGATED LAND USE TYPES INCLUDED IN THE LAND USE CHANGE MODELLING, THE AREA OF THESE LAND USE TYPES, THE PROPORTION OF PASTURE AND CROPLAND IN THESE LAND USE TYPES AND THE PROPORTION OF CROP AND PASTURE DEMAND THAT IS MET BY THE SPECIFIC LAND USE TYPES IN 2005. TABLE 4.6: CHARACTERISTICS AND WEIGHTS OF THE SUITABILITY FACTORS FOR LAND USE ALLOCATION. TABLE 4.7: CATEGORIES OF LAND USE TYPES, LAND COVER TYPES AND PHYSICAL CONSTRAINTS EXCLUDED FROM

109

110 112 127

150

151 152

154 156

LAND USE CHANGES TOWARDS CROPLAND AND PASTURE AND EXCLUDED FROM THE LAND AVAILABLE FOR BIOENERGY CROPS. TABLE 4.8: STOCHASTIC INPUT VARIABLES. TABLE 5.1: LAND AVAILABILITY FOR BIOENERGY CROPS IN MOZAMBIQUE IN THE TIMEFRAME 2010-2030 FOR THE BUSINESS AS USUAL AND THE PROGRESSIVE SCENARIO. TABLE 5.2: CULTIVATION COST ITEMS OF EUCALYPTUS AND SUGARCANE IN MOZAMBIQUE. TABLE 5.3: MAXIMUM AND AVERAGE YIELD LEVELS OF EUCALYPTUS AND SUGARCANE IN MOZAMBIQUE UP TO 2030 GIVEN THE LAND AVAILABILITY IN THE BUSINESS AS USUAL AND THE PROGRESSIVE SCENARIO. TABLE 5.4: CONVERSION PLANT CHARACTERISTICS FOR WOOD/TORREFIED PELLETS AND ETHANOL PRODUCTION FOR SHORT (2010) AND LONG (2030) TERM. TABLE 5.5: TRANSPORT COST IN €/TONKM FOR SEVERAL ROAD TYPES AND CONDITIONS. TABLE 5.6: TRANSPORT AND STORAGE CHARACTERISTICS OF WOOD PELLET AND ETHANOL FOR ROAD TRANSPORT AND INTERNATIONAL SHIPPING. TABLE 5.7: RANGES FOR SENSITIVITY ANALYSIS OF TOTAL COSTS OF SUPPLY CHAIN FOR EUCALYPTUS AND SUGARCANE IN 2030 IN THE PROGRESSIVE SCENARIO. TABLE 6.1: MAIN CHARACTERISTICS OF THE TWO SCENARIOS RELATED TO LAND USE CHANGE TABLE 6.2: SUITABILITY FACTORS FOR LAND USE ALLOCATION OF CROPLAND, MOSAIC CROPLAND-PASTURE AND PASTURE. TABLE 6.3: SPATIAL DATA INPUT FOR PLUC MODEL, THE DATA SOURCE OF THE REQUIRED DATASETS, THEIR RESOLUTION AND THE ADAPTATIONS MADE TO THESE DATASETS. TABLE 6.4: INPUT PARAMETERS FOR CALCULATIONS OF CHANGES IN CARBON STOCK FOR THE BAU EN PROGRESSIVE SCENARIO AND THE PROGRESSIVE SCENARIO WITH MITIGATION MEASURES.

XV

158 160 167 170 171 172 176 177 188 209 213 214 240

TABLE 6.5: EMISSION FACTORS OF DIRECT AND INDIRECT N2O EMISSIONS DIFFERENTIATED FOR SEDIMENT TYPE, LAND USE AND NITROGEN SOURCE. TABLE 6.6: ABATEMENT FIGURES FOR THE REPLACEMENT OF GASOLINE BY ETHANOL FROM SWITCHGRASS AND WHEAT BASED ON THE FIGURES PROVIDED BY JRC. TABLE 7.1: OVERVIEW OF THE SETTINGS OF THE THESIS CHAPTERS AND THE RESEARCH QUESTIONS ADDRESSED IN THEM. TABLE 7.2: OVERVIEW OF THE POTENTIALS, COST AND ENVIRONMENTAL IMPACTS OF BIOENERGY PRODUCTION IS PROVIDED FOR THE DIFFERENT GEOGRAPHICAL SETTINGS.

XVI

241 242 249 260

Units and Abbreviations AEZ

Agro-ecological zoning

EU

European Union

BAU

Business as Usual

exp.

Exponential

C

Carbon

FOB

Freight on Board

Ca

Calcium

GDP

Gross Domestic Product

CAP

Common Agricultural Policy

GHG

Greenhouse gas

cap

Capita

GIAB

CAPEX

Capital Expenditures

Geografisch Informatiebestand Agrarische Bedrijven

Cc

Carbon content

GIS

Geographic Information System

CEEC

Central and Eastern European Countries

GJ

Giga joule, 1 GJ = 10 Joule

CH4

Methane

GT

Giga ton, 1 GT = 10 ton

CLUE

Conversion of Land Use and its Effects

GWP

Global Warming Potential

h

hour

CO2

Carbon dioxide

ha

hectare

CO2-eq

Carbon dioxide equivalent

HELP

Her-Evaluatie Landinrichtings Project

DDGS

Dried Distillers Grains and Solubles

HHV

Higher Heating Value

DEM

Digital Elevation Map

HI

Harvest index

DM

Dry Matter

HNV

High Nature conservation value

EC

European Commission

iLUC

Indirect Land Use Change

EF

Emission factor

IMAGE

Integrated Model to Assess the Global Environment

EJ

Exa joule, 1 EJ = 10 Joule

inv. prop. inverse proportional

EP

Effective precipitation

IPCC

ES

Additional sugar beet

Intergovernmental Panel on Climate Change

ET0

Reference soil organic Carbon level

ISO

International Organisation Standardisation

18

XVII

9

9

for

K

Potassium

MSA

Mean Species Abundance

K2O

Potassium Oxide

Mt

Megaton, 1Mt = 10 kg

KC

Crop evapotranspiration coefficients

MW

Megawatt, 1MW = 10 watt

kcal

kilo calorie, 1 Kcal = 10 Kcal

MZN

Mozambican Metical

kha

kilo hectare

N

Nitrogen

Km

kilometer

N2O

Nitrous Oxide

square kilometer

NH3

Ammonia

km

2

3

6

9

6

kton

kilo ton, 1 kt = 10 kg

NL

The Netherlands

kW

kilowatt

no.

number

kWh

Kilowatt-hour

NO3

Nitrate

l

litre

NOx

Nitrogen Oxide

LCA

Life Cycle Analysis

NPV

Net Present Value

LDC

Least Developed Countries

NUTS

LHV

Lower Heating value

Nomenclature of Units for Territorial statistics

LUC

Land Use change

O&M

Operation and Maintenance

m

meter

odt

oven dried ton

M€

Million Euro

P

Phosphorus

cubic meter

p.a.

m

3

MC Mg mg

-

Monte Carlo

p/km 6

Mega gram, 1 Mg = 10 gram -3

Milligram, 1mg = 10 gram 6

per annum 2

persons per square kilometre

P2O5

Phosphorus pentoxide

PE

Effective precipitation

PJ

Peta Joule, 1 PJ = 10 Joule

15

Mha

Mega hectare, 1Mha = 10 hectare

MJ

Mega joule, 1MJ = 10 joule

PLUC

PCRaster Land Use Change

mln

Million

POUL

Poultry

mm

millimeter, 1mm = 10 m

ppm

parts per million

Moz

Mozambique

PROG

Progressive

r

scale factor

6

-3

Mpeople Million people

XVIII

s

second

TC

Ton Cane

SA

South Africa

th(in)

thermal input

SADC

Southern African Development Community

TOP

Torrefied Pellet

UKR

Ukraine

SD

Standard Deviation

US$

United States Dollars

SGe

Small grain equivalent

VAT

Value Added Tax

SOC

Soil Organic Carbon

veg

vegetables

SOCref

Reference soil organic Carbon level

WEC

Western European Countries

SOM

Soil Organic Matter

WEQ

Wind Erosion Equation

SSF

Simultaneous Saccharification and Fermentation

WUE

Water Use Efficiency

y

year

SSR

Self Sufficiency ratio

Yld

Yield

t

Metric tonne

XIX

1.

Introduction

1

Introduction

21

SHADES OF GREEN

1.1 Current and future energy supply Society requires energy services to meet basic human needs and to serve productive processes (IPCC 2011). Current global primary energy supply amounts to nearly 500 EJ (2008) and is expected to increase to between 600 and 1000 EJ in 2050 (IPCC 2007b). This increase is primarily driven by population growth and an increasing economic welfare. Population growth is projected to be 0.4-1.1 % per year, resulting in 8 to 10 billion people by 2050 (UNDP 2011); while economic welfare increases in parallel with an estimated average annual gross domestic product (GDP) growth of 3.5% (IEA and OECD 2011). Currently 85% of the total primary energy supply originates from fossil fuels (IEA 2010a; IPCC 2011). This fossil fuel dominated energy supply cannot be sustained in the long run for a number of reasons. Although, the size and the accessibility of fossil energy sources (oil, coal and gas) remain a point of discussion, the easily accessible reserves will be exhausted eventually (UNDP et al. 2000). Also, the reliance on fossil fuels increase dependency on politically unstable areas and therefore negatively affects energy security (UNDP et al. 2000; UNDP et al. 2004). Moreover, fossil fuels are the most important contributor (56%) to anthropogenic greenhouse gas (GHG) emissions and are therefore a major cause of climate change (IPCC 2007b; 2011). Atmospheric concentrations of greenhouse gases, exceeded 390 parts per million (ppm) in CO2- equivalent in 2010. This is 39% higher than pre-industrial levels and emission rates continue to increase (IPCC 2011). The uneven geographical distribution, the exhaustion of supplies of fossil fuels, and negative external effects, are likely to result in higher and more variable costs of energy production, hampering sustainable economic development. For all those reasons, a transition towards a more sustainable energy system is required.

1.2

The role of bioenergy

There are several options to make the energy system environmentally and economically more sustainable. A precondition is the decoupling of population growth and increasing welfare on the one hand and an increasing energy demand on the other. This could be established by means of reduced consumption and higher energy efficiencies. In addition, CO2 emissions can be reduced via deployment of nuclear energy and via fossil fuel conversion combined with carbon capture and storage equipment and deployment. However, increased deployment of renewable energy resources to substitute of fossil fuels is considered essential to make the energy system sustainable. Currently, renewable energy sources contribute 13% to the total primary energy supply (IEA 2010a), of which almost 80% (50 EJ) is supplied by biomass (IEA 2010a). Approximately 60% of this biomass is used in traditional ways, such as the use of wood and charcoal for cooking and heating by the poorer part of the population in developing countries (IPCC 2011). This traditional 22

1.

Introduction

use of biomass has many disadvantages such as the low energy efficiency, the labour intensity of fuel wood collection, indoor air pollution and related health problems, and deforestation and soil degradation (Karekezi 2002; Bruce et al. 2006; IEA 2010b). Modern biomass applications such as the use of biomass for electricity, heat and biofuel production contributed 11 EJ in 2008 (IPCC 2011). Modern use of biomass is steadily increasing and is expected to play an important role in future energy supply (IEA and OECD 2011; IPCC 2011). Recent literature has shown that there is a large technical bioenergy potential (IPCC 2011), which allows for an increasing contribution of bioenergy to the total primary energy supply. For example, Dornburg et al. (2010) calculate a technical bioenergy potential of 500 EJ per year. In the Special Report on Renewable Energy Sources and Climate Change Mitigation of the Intergovernmental Panel on Climate Change (IPCCSRREN), the global deployment potential of energy from biomass in 2050 is estimated to amount between 100 – 300 EJ, a two to six-fold increase compared to current total (traditional + commercial) biomass use (IPCC 2011). The increasing production and use of bioenergy is driven by several key factors. First, the substitution of fossil fuels with biomass could contribute substantially to mitigate global GHG emissions, under the condition that feedstocks are produced sustainably and that efficient bioenergy systems are used (Dornburg et al. 2010; IPCC 2011). Second, biomass can be converted to heat, power, (liquid and gaseous) transport fuels and used as feedstock via a range of technologies and deployed relatively easily in existing energy infrastructure. Particularly in the transport sector (such as aviation and truck transport), there are few alternative renewable energy sources that could be implemented without far reaching modifications to the current vehicles and energy infrastructure. In addition, biomass could be used for other applications to substitute fossil resources such as the production of biochemicals and biomaterials. Especially biorefining of biomass for several applications such as food, feed, fibre, energy and chemicals offers opportunities for efficient use of biomass resources (Sanders and van der Hoeven 2008). Third, biomass resources are geographically more evenly distributed which could diversify energy supply and contribute to an improved energy security. Fourth, biomass and bioenergy production can contribute to rural development by providing a stable and additional income for example, through export of biomass-derived commodities for the world’s energy market (IPCC 2011). Fifth, herbaceous or woody energy crops used for re-vegetation of degraded land can lead to the restoration of the land and could have positive environmental and socio-economic impacts e.g. increased soil organic carbon, economic benefits from soil restoration, and increased biodiversity (Wicke et al. 2012). Dedicated bioenergy crops are considered to become the main contributors to future bioenergy supplies (WWI 2006; Smeets et al. 2007; Dornburg et al. 2010) if higher 23

SHADES OF GREEN

deployment levels are achieved. However, an increased implementation of dedicated bioenergy crop production could have significant adverse socio-economic and environmental impacts such as deforestation, loss of carbon sinks, biodiversity and other ecosystem functions and services, displacement of people, increased competition with food and higher food prices (IPCC 2011; Wicke et al. 2012). The latter is especially detrimental for developing countries. Many of these impacts are related to land use change (LUC) (Wicke et al. 2012) which could either directly or indirectly be caused by the additional land use for bioenergy production. Direct land use change refers to a situation in which previous land use is converted to energy crop production. Indirect land use change (iLUC) is land use change which occurs outside the production boundary of a feedstock, but which is caused by a change in the use or level of output of that feedstock (Tipper et al. 2009). Therefore, expansion of bioenergy in the absence of monitoring and good governance of land use carries the risk of significant conflicts with respect to food supplies, water resources and biodiversity, as well as a risk of low GHG benefits. In order to achieve the high potential deployment levels of biomass for energy, competition between food, feed and fuels - and therefore also indirect land use change (iLUC), need to be avoided by balancing the increased production of biomass for energy by improvements in agricultural management (Dornburg et al. 2010; Wicke et al. 2012). Furthermore, the key environmental concerns should be addressed by selecting appropriate bioenergy systems and applying adequate land use planning (Dornburg et al. 2010). The impacts and performance of biomass production and use are region- and sitespecific (IPCC 2011) and impacts occur at micro to macro scales (van Dam et al., 2010). Implementation of effective sustainability frameworks, for example, by developing certification schemes, could mitigate negative environmental impacts and allows simultaneously for contributing to multiple objectives of sustainable development. Figure 1.1 depicts the complex interactions of bioenergy with the socio-economic and environmental context at micro and macro scaleas they were illustrated by the IPCC (IPCC 2011). The risks are generally related to business-as-usual approaches with uncoordinated implementation of large scale bioenergy, and the opportunities could generally be achieved provided good governance and sustainability frameworks are implemented (derived from IPCC 2011). At several levels, initiatives for sustainability criteria, codes of conduct and protocols have been and are currently developed to deal with these issues (e.g. Projectgroep 'Duurzame productie van Biomassa' 2006; Fehrenbach et al. 2008; Gallagher 2008; ISO 2008; EC 2009a; NEN 2009; ISCC 2010; RSB 2010; GBEP 2011). Sustainability criteria include general principles related to legality, economic viability, and environmental and socio-economic impacts. At present, a key bottleneck for both market players and government is how such 24

1.

Introduction

criteria can be met in practice and how impacts can be quantified in a verifiable and reliable manner. As impacts and performance of bioenergy production interacts with a dynamic socio-economic and environmental context, at local, regional and global level, integrated analysis at different spatial and temporal scales is required for coherent full impact analysis of biomass production and use. On various key aspects (such as biodiversity, socio-economic parameters, avoidance of iLUC and even methods for calculation of GHG balances) there is limited to no agreement on the scientific methods to determine impacts of biomass production (Smeets et al. 2008; van Dam et al. 2008). Strong improvement in spatially explicit potential and impact analyses are desired to allow for advanced certification, sound planning of future and sustainable investments and good governance of land-use and the agricultural sector, in direct relation to increasing biomass production and use for energy and materials.

Figure 1.1: Complex interactions of bioenergy with the socio-economic and environmental context at micro and macro scale, taken from IPCC (2011).

1.3 Need for regional specific assessments Land availability and the yield that can be obtained on available land are two critical determinants of the technical and sustainable potential of bioenergy crop production (IPCC 2011). In recent years, an increasing number of studies have been published on 25

SHADES OF GREEN

bioenergy potentials on a global (e.g. Berndes et al. 2003; Hoogwijk et al. 2005; Smeets et al. 2007; Hoogwijk et al. 2009; Dornburg et al. 2010; Beringer et al. 2011), European (e.g. Ericsson and Nilsson 2006; EEA 2007; Fischer et al. 2007; de Wit and Faaij 2010), national (e.g. Faaij et al. 1998; van den Broek et al. 2001; Walsh et al. 2003; Perlack et al. 2005; Sang and Zhu 2011) and regional level (e.g. van Dam et al. 2009a); taking into account various levels of constraints related to environmental protection. However, most of these studies have assessed biomass potentials on a spatially aggregated level. A major disadvantage of such studies is that they provide limited up to no information on the location of the land that is or could become available for bioenergy crops and do not account for high spatial heterogeneity in yield levels and environmental impacts. The economic performance of bioenergy production depends on the cost of biomass feedstock, the costs of transport and the cost and the efficiency of the conversion technology. Feedstock costs are often a dominating cost factor in the total costs of energy, which are in turn dependent on biomass crop yield, cost of production factors (such as fertilizer, labour and land), and supply chains of the produced biomass. Several studies on biomass potentials mentioned above include cost estimations. Because of the spatially aggregated level of these studies, regional averages of yield, and transport distances were assumed which mask the significant spatial variability in cost of production. Because of developments in land availability, costs of inputs and technological learning, the economic performance of bioenergy are dynamic. Including technological learning in the estimation of future cost of bioenergy costs have been demonstrated by (e.g. Junginger et al. 2006; van den Wall Bake et al. 2009; de Wit et al. 2011b). The environmental impacts of bioenergy production related to GHG emissions, and impacts on soil water and biodiversity are strongly dependent on the specific biophysical conditions such as previous land use, soil characteristics, precipitation, slope, etc. of the production region. However, many studies on the environmental impacts of bioenergy production use a generic LCA approach (Hamelinck and van den Broek 2005; Blottnitz and Curran 2007) and/or use national or regional averages of biophysical factors such as soil type, precipitation levels and land use (e.g. Smeets and Faaij 2010; de Vries et al. 2011b). Van Dam et al (2009b) deployed an ex ante socio-economic and environmental impact analysis of energy crops at a regional level. They found large variations in the performance of energy crops for various impacts for the different settings (including reference land use, agro-ecological suitability, crop type and management applied). De Vries et al. (2011a) assessed the environmental impacts of three energy crops for different locations and found considerable differences in performance between crops and locations. Also, several studies focus only on one specific aspect of environmental performance such as GHG (e.g. Kim et al. 2009; Hoefnagels et al. 2010), water (e.g. 26

1.

Introduction

Berndes 2002; Gerbens-Leenes et al. 2009) or biodiversity (e.g. Semere and Slater 2007; Sala et al. 2009). However, many studies show that there are several tradeoffs between environmental impacts (e.g. van Dam et al. 2009b; De Vries et al. 2011a; de Vries et al. 2011b). Therefore, the environmental impacts of bioenergy production should be assessed spatially and in an integrated way. In agricultural science, an increasing number of studies have assessed the environmental impacts of the agricultural sector spatially explicitly (e.g.Leip et al. 2008; Lesschen 2008; Smith et al. 2008; Velthof et al. 2009; Verburg et al. 2009). Lapola et al.(2010) conducted a spatially explicit assessment of potential iLUC induced by a fixed bioenergy implementation target in Brazil, and calculated the associated GHG emissions. Some studies demonstrated methods to assess the GHG impacts of the implementation of bioenergy production and the agricultural sector in an integrated way (Melillo et al. 2009; de Wit et al. 2011c; Popp et al. 2011). However, because of their geographical scope (Global, European), these studies were conducted at a spatially aggregated level.

1.4

Aim and thesis outline

Based on the gaps of knowledge identified in existing literature, the main aim of this thesis is to examine how potentials, costs, and environmental impacts of bioenergy production can be assessed, taking into account the avoidance of ilUC and the spatiotemporal variability of the biophysical and socio-economic context. Therefore, the following research questions are addressed: I. How can potential land availability for energy crops be assessed spatially and temporal explicitly, given that iLUC should be avoided and therefore taking into account also the development in other land use functions? II. How can the economic viability (the location specific competition with other agricultural land uses, the cost of biomass feedstock production and the logistics of the supply chain) and the environmental impacts of bioenergy production (impacts on GHG emissions, soil, water and biodiversity) be assessed spatially and temporally explicitly. III. What are the potential land availability, economic performance and environmental impacts of bioenergy production in different settings? IV. What reliability can be obtained using the data available and the methods developed in this study? The research questions are addressed in Chapters 2 through 6. In Chapter 2 and 3, research question I is addressed by assessing the economic viability and the potential 27

SHADES OF GREEN

environmental impacts of regional bioenergy production chains spatially explicitly, taking into account spatial variation in agro-ecological suitability, current land use and other biophysical factors. The methodologies developed in these chapters allow for spatial explicit but static assessments of costs and environmental impacts. The north of the Netherlands was selected as a case study area because of the high competition for land related to the intensive land use and because of the good data availability. In Chapter 4, a new land use change model was developed to assess the land availability for bioenergy crops, taking into account the development in other land use functions such as cropland, pasture and forest. The model developed allows for spatial and temporal explicit assessment of bioenergy potentials. Subsequently, the developments in costs of biomass feedstock and the logistics of bioenergy supply chains were assessed, given the developments in potential land availability and technological learning in chapter 5. The coupling of the spatiotemporal land use model, the spatially explicit biomass production and logistics costs, and the temporal cost developments, allowed for spatiotemporal assessment of bioenergy supply costs. In chapter 6 the developed land use change model was adapted for Ukraine and extended with GHG emission module in order to analyze the potential developments in the GHG emissions of the entire agricultural sector, including the implementation of bioenergy crops and an intensification of the agricultural sector. The coupling between a dynamic GHG emission calculation module to the land use change model allows for spatiotemporal GHG impact assessment. Mozambique and Ukraine were selected as case study areas because of their high bioenergy production potential related to the low population density and the favourable climate for biomass production, and because of the diversity of the environmental and socio-economic conditions of the countries. In chapters 2, 3, 5, and 6 the performances of typical first and second generation energy crops, (mostly) assumed to be used for ethanol production, to allow for a comparison of the differences in potentials, costs, and environmental impacts for typical first and second generation biofuel options. The complexity and level of integration of methodologies developed in this thesis increase from chapter to chapter and evolves from spatially explicit and static modelling (in Chapter 2 economic performance and in Chapter 3 environmental impacts) to spatiotemporal and dynamic modelling (in Chapter 4 land use change and in Chapter 5 cost supply development). In chapter 6, the spatiotemporal land use modelling is integrated with dynamic GHG emission modelling, which allows for spatiotemporal integrated impact assessment. In all chapters, the limitations of the methods and the data used in the assessments are discussed. Table 1.1 presents an overview of the chapters and the research questions addressed.

28

1.

Introduction

Table 1.1: Overview of the topics of the thesis chapters and the research questions addressed in them. Chapter

Topic

2

Potential, spatial distribution and economic performance of regional biomass chains Spatial variation of environmental impacts of regional biomass chains Spatiotemporal land use modelling to assess land availability for energy crops Spatiotemporal cost-supply curves for bioenergy production Integral spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances

3 4 5 6

Research questions I II III •































IV

Chapter 2 addresses research question II, III and IV by analyzing the spatial variation in economic viability of ethanol production from Miscanthus and sugar beet in the North of the Netherlands. The competitiveness of bioenergy crops is analysed by comparing the economic performance of perennial crops, current rotations, and rotation schemes which include additional years of sugar beet and by comparing the production cost of bioethanol with average petrol prices. The current land use and soil suitability for present and bioenergy crops are mapped using a geographical information system (GIS) and the spatial distribution of economic profitability is used to indicate where land use change is most likely to occur. Chapter 3 addresses research questions II, III and IV by quantitatively assessing the spatial variation of potential environmental impacts of ethanol production from Miscanthus and sugar beet in the North of the Netherlands. The environmental impacts included are impacts on greenhouse gas (GHG) emissions (during lifecycle and related to direct land use change), soil, water, and biodiversity. Suitable methods are selected and adapted based on an extensive literature review. The spatial variability of the environmental impacts in relation to previous and current land use, soil characteristics, climate conditions, crop characteristics, management etc is assessed using Geographical Information System (GIS). Chapter 4 addresses research question I, III and IV by developing a new land use change model to assess future developments in land availability for bioenergy crops in a spatially explicit way, while taking into account the developments in other land use functions, such as land for food, livestock and material production. The land use dynamics are modelled based on the suitability of a location for a specific land use related to the agro-ecological suitability, accessibility, conversion elasticity and neighbourhood relations. This spatiotemporal land use change model (PLUC) developed allows for uncertainty assessment by stochastic modelling of the key land use change drivers and suitability 29

SHADES OF GREEN

factors. The model (PLUC) is demonstrated with a case study on the developments in the land availability for bioenergy crops in Mozambique within the period 2005-2030. Chapter 5 addresses research questions II, III and IV, by assessing how bioenergy supplies, production and supply costs develop spatially explicitly over time. Focus lays on export oriented biomass and biofuel production. This analysis is based on the developments in land availability (resulting from the analysis of chapter 4), the suitability of the land that is could become available, the disaggregated cost break down of energy crop production, the transportation distance of feedstock to conversion plant, the cost of conversion, the transportation distance from plant to harbour and the cost of international shipping. Moreover, the cost reductions in biomass feedstock production and conversion achieved by technological learning over time are taken into account. The supply chains of eucalyptus (torrefied) pellets and sugarcane ethanol in Mozambique are used as a case study. Chapter 6 addresses research question I, II, III and IV by analysing spatially explicitly how bioenergy potential and greenhouse gas (GHG) emission reduction in Ukraine can develop in the timeframe 2010-2030, while taking into account development and emissions of other agricultural land use functions at the same time. The development in land requirements for food and feed production is analyzed spatially explicitly on an annual basis making use of the land use change model PLUC. The results of the PLUC model provide information on time and location of land use and management changes. The land use dynamics are the input for the spatiotemporal GHG emission assessment module which includes calculations of CO2, N2O and CH4 emission related to changes in agricultural management and land use and the abatement of GHG emissions by replacing fossil fuels by bioethanol produced from wheat and switchgrass. Chapter 7 summarises and evaluates the findings from chapter 2 to 6, provides answers to the research questions, highlights the possibilities and limitations of the methods and data used, and gives recommendations for further research.

30

32

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

2

Potential, spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

F. van der Hilst, V. Dornburg, J.P.M. Sanders, B. Elbersen, A. Graves, W.C. Turkenburg, H.W. Elbersen, J.M.C. van Dam, A.P.C. Faaij Agricultural Systems (2010) Volume 103, Issue 7: 403-417

33

SHADES OF GREEN

ABSTRACT This work assesses the viability of regional biomass chains by comparing the economic performance of potential bioenergy crops with the performance of current agricultural land uses. The biomass chains assessed are ethanol production from Miscanthus and from sugar beet in the North of the Netherlands. The competitiveness of bioenergy crops is assessed by comparing the Net Present Value (NPV) of perennial crops, current rotations, and rotation schemes, which include additional years of sugar beet. The current land use and soil suitability for present and bioenergy crops are mapped using a geographical information system (GIS) and the spatial distribution of economic profitability is used to indicate where land use change is most likely to occur. Bioethanol production costs are then compared with petrol costs. The productions costs comprise costs associated with cultivation, harvest, transport and conversion to ethanol. The NPVs and cost of feedstock production are calculated for seven soil suitability classes. The results show that at current market prices, bioenergy crops are not competitive with conventional cropping systems on soils classified as “suitable”. On less suitable soils, the return on intensively managed crops is low and perennial crops achieve better NPVs than common rotations. Our results showed that minimum feedstock production costs are 5.4 €/GJ for Miscanthus and 9.7 €/GJ for sugar beet depending on soil suitability. Ethanol from Miscanthus (24 €/GJ) is a better option than ethanol from sugar beet (27 €/GJ) in terms of costs. The cost of bioethanol production from domestically cultivated crops is not competitive with petrol (12 €/GJ) production under current circumstances. We propose that the method demonstrated in this study provides a generic approach for identifying viable locations for bioenergy crop production based on soil properties and current land use. 34

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

2.1 Introduction Energy from biomass, including biofuels like ethanol, can play a major role in local, national and global energy supplies depending on land availability, costs, and supply. However, in both scientific and political arenas, it is seen that such bioenergy chains need to evolve in a way that is compatible with sustainable development. In recent years, several studies (e.g.Hoogwijk et al. 2005; Smeets et al. 2007; Dornburg et al. 2008) have assessed the world bioenergy potential and the contribution to the world energy demand. Other studies have focused on bioenergy potential and related costs at a European level (EEA 2006; van Dam et al. 2007; de Wit and Faaij 2010; Fischer et al. 2010a; Fischer et al. 2010b) or national level (e.g. van den Broek et al. 2001; Batidzirai et al. 2006; Styles and Jones 2007a). However, few studies describe the spatial variation of bioenergy production potential and the cost of bioenergy supply within a region. Since the physical environment is spatially heterogeneous, location is a key factor for the economic viability and environmental performance of bioenergy production. Because economic benefit is a major incentive for adoption, this paper focuses on the competitive advantage of bioenergy crops in relation to conventional land use in order to increase understanding of where, and on which types of soils, such land use changes might occur. Ethanol production from Miscanthus (Miscanthus x Giganteus) and sugar beet (Beta vulgaris L.) in the North of the Netherlands is selected for our case study. This region is important as a test case, because of the high pressure on land for various uses including intensive agriculture. This enables an extensive analysis of the economic viability of regional biomass chains. Sugar beet and Miscanthus are selected because of their high potential yields and because they represent a typical first and second generation bioenergy chain. These are compared with current land use to determine their relative economic viability. In Section 2.2 we elaborate on the design of the bioenergy chains, the characteristics of the region and the potential land availability in the region. In Section 2.3, the methods applied to assess the competitiveness of new bioenergy crops compared to current land use and the methods to calculate the cost of feedstock and ethanol production are discussed. The approach to determine the soil suitability and the effect on the spatial variation of economic performance of potential and current land use is described in section 2.3.3. In Section 2.4 the results of the assessment are presented and the spatial variation is depicted in maps of the region. A sensitivity analysis shows the level of robustness of the results. In Section 2.5, the applied method, the data used and the results are discussed, and in Section 2.6, conclusions are drawn. 35

SHADES OF GREEN

2.2 Case study description 2.2.1 Study region The Northern region of the Netherlands (Groningen, Friesland and Drenthe) was selected as the area for our research for several reasons. Firstly, the Dutch government has provided clear targets for substitution of fossil fuel and green house gas emission reduction (Menkveld 2007; Ministerie van VROM 2007; Ministerie van Economische zaken 2008). Secondly, the pressure on land is relatively high due to a high population density, diverse land uses and an intensive agricultural sector, resulting in intense competition between different land uses. Thirdly, access to sea transport through the Eemshaven ports facilitates the possible transport of biomass feedstock and intermediate- or end-products to and from the rest of the world. Fourthly, this is a highly productive agricultural area with fertile soils, favourable climatic conditions, and advanced agricultural management (Romkes and Oenema 2004) with a farming population that is interested in alternative economic activities for the agricultural sector. Finally, several regional stakeholders have also articulated on the need for sustainable development in the region (Costa Due 2009; Energy Valley 2009). The region has a mild maritime climate with average temperatures of 16 ºC during summer and 3ºC during winter (KNMI 2002). The most common soil types in the Northern region of the Netherlands are sand, clay, sandy clay and peat, and soils are generally fertile. Precipitation is relatively high as are ground water levels. The climate and soils are suitable for a wide range of crops (Christian et al. 2001). Land use in the region (1.1 Mha) is dominated by agricultural activities: 68% of the total area is agricultural land of which 41% is used for agricultural crops and 57% for pastures. On parts of the pasture areas, silage maize is continuously cultivated by intensive cattle breeders. Cereals, potatoes, sugar beet and silage maize are the most dominant crops cultivated in rotation. Two common rotations schemes for sandy soils and two rotations schemes for clay soils are selected to represent current land use of arable land in the region and are depicted in Table 2.1. Transport infrastructure in the region is well developed. Whilst waterways and railways are available, road transport is the most convenient way of transporting agricultural goods within the region due to the relatively short distances and the flexibility that multiple production sites require (Hamelinck et al. 2005b). Rail and waterways and the Northern ports, connect this region to the rest of Europe and beyond.

36

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

Due to intensive livestock production, the Netherlands faces a manure surplus. Because of the costs of managing this surplus, the application of manure on agricultural land has negative costs. Therefore, application rates are high in pasture areas with intensive cattle breeding. Table 2.1: Two typical rotation schemes for sandy soils and two typical rotation schemes for clay soils for Northern region of the Netherlands derived from (LEI CBS 2007; van der Voort et al. 2008) expressed as proportion of individual crop in each of the rotations. Share of crop in rotation Winter wheat Summer barley Winter barley Seed potato Industrial potato Sugar beet Maize Other Fallow Total

Clay rotation

Sand rotation

I

II

I

0.57

0.20 0.10

0.28 0.06 0.03 0.30 0.20 0.04 0.06 0.04 1.00

0.20

0.14

0.09 1.00

0.15 0.15 0.10 0.25 0.05 1.00

II 0.05 0.25 0.05 0.45 0.20

1.00

2.2.2 Biomass potential in the region The introduction of bioenergy crops to large areas of land would create competition with the food and feed crops already being grown in the region. Thus, in order to define a limit to the arable land available for bioenergy production, information provided by the EU Refuel project is used (de Wit and Faaij 2010). One of the objectives of the Refuel project was to map the potential production and costs of biomass crops in the EU for different time frames and for several land use scenarios. The method used in this study is comparable with the approaches used by Smeets et al. (2007) and van Dam et al (2007). In the Refuel approach, projections are used to describe the future dynamics of population growth, food intake per capita, agricultural production intensity, livestock intensity and land requirements for the growth of cities, villages and infrastructure (Fischer et al. 2010a; Fischer et al. 2010b). The land available for biomass production is calculated by subtracting the land needed for other land use functions (including nature) from the total available land, assuming the self-sufficiency in food production in the region remains constant. In the Refuel study it is assumed that typical agricultural crops are only produced on arable land, while for herbaceous crops like Miscanthus it is assumed that pasture could also become available.

37

SHADES OF GREEN

The base case scenario of the Refuel assessment is derived from the Common Agricultural Policy (CAP) of the EU. In addition, a more optimistic (high land availability) and a more pessimistic (low land availability) variant have been developed. In Table 2.1, the share of agricultural land that according to the Refuel results could become available for biomass production in the North of the Netherlands in 2015 and 2030 is depicted. The Refuel projections of land availability for biomass production in the North of the Netherlands are somewhat higher but in the same order of magnitude as the projected land availability of the Eururalis project (Westhoek et al. 2006; Eickhout and Prins 2008). In this study, the Refuel project is used to indicate what proportion of land could be converted for bioenergy production without diminishing the region’s current selfsufficiency in food. In addition, data from the Refuel project are used to estimate the appropriate scale of conversion plants for the region. Table 2.2: Proportion of land that could become available for biomass production in North of the Netherlands according to three Refuel scenarios. Type of land

Arable Pastures

Availability in % of land Low 2015 2030

Medium 2015 2030

2015

1.9 0.5

2.7 0.5

4.3 0.5

6.1 8.6

7.4 8.6

High 2030 10.2 8.6

2.2.3 Bioenergy chains In this study, we investigate ethanol production from sugar beet and Miscanthus. These two bioenergy chains are selected because of their potential for high yields (Huisman et al. 1997; Elbersen et al. 2005; de Wolf and van der Klooster 2006; van der Voort et al. 2008) and because of the developing market for ethanol in the Netherlands created by European biofuel policies. In addition, the two bioenergy chains are chosen because they have very different cultivation requirements and conversion technologies, since they are typical of first and second generation bioethanol supply chains. Sugar beet Sugar beet requires good quality soils and high inputs and is generally grown in rotation with cereals and potatoes. In our study, it is assumed that sugar beet for ethanol production is cultivated on land currently in use as arable land (as in the Refuel study, pasture is excluded for typical agricultural crops). This implies that the proportion of sugar beet needs to be increased within the current rotation schemes. Because the excessive use of beet or other intensive crops increases the risks of diseases and yield loss (Kempenaar et al. 2003), it is assumed that the proportion of sugar beet does not exceed 38

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

25% of the rotation, and that the total proportion of land assigned to intensively managed crops does not increase from current levels. Current CAP regulations for sugar include a quota and a price regime. The quota limits the production of sugar per county and the price regime sets a guaranteed intervention price for this quota. Sugar produced over the quota is sold on the world market, at considerably lower prices than EU quota prices. Since extra beet exceeds the quota for sugar production, sugar beet for ethanol production is less profitable than for sugar production. For this reason, it is assumed that the growth of sugar beet for ethanol is additional to that sold as the sugar beet quota. Management and transport of sugar beet for ethanol production is assumed to be similar to current practice in this region. Once harvested, sugar beet cannot be preserved. The harvest window lasts from September until the end of December, thus maximizing the load factor of the beet processing plant. It is assumed the sugar beet, including 15% tare (soil attached to the beet), is transported by truck to a newly built ethanol plant close to the current sugar plant centrally located within the agricultural area. Since long distance transport of sugar beet is not economically attractive, the conversion plant is assumed to be of a size appropriate for the expected supply of sugar beets in the region, i.e. 700 kton (fresh weight) input per year (90 MWinput, 1.5 PJethanol). This figure is derived from predictions made in the Refuel project on the maximum land available for arable bioenergy crops in 2015 (9.6 kha, see section 0) and -1 -1 the attainable yield on very suitable soils (73 tonfresh ha y , 23% DM, 16% sugar). In the ethanol plant, sugar beet is shredded into cossettes and diffused in water to produce raw sugar beet juice and pulp. Pulp is further processed for animal feed and put on the market as a co-product. The raw juice is pasteurized, fermented, and distilled in order to produce ethanol. Miscanthus Miscanthus is a perennial crop with a rotation of 20 years. It requires few inputs and is relatively insensitive to soil conditions (Venturi et al. 1999; Bullard 2001; Bullard and Matcalfe 2001; Lewandowski and Heinz 2003; Lewandowski et al. 2003; Khanna et al. 2008). In our study, it is assumed that Miscanthus can be cultivated on agricultural land that is currently in use as arable land and as pasture (as in de Refuel study de Wit and Faaij 2010). Although the highest yields are achieved when Miscanthus is harvested in autumn, harvest does not take place until spring, when the highest dry matter content and quality is achieved. Due to nutrient remobilization during winter, the removal of nutrients from the soil is lower in delayed harvests (Himken et al. 1997; Ercoli et al. 1999; Lewandowski and Heinz 2003; Monti et al. 2008) and this is preferable, since lower moisture, nutrient 39

SHADES OF GREEN

and ash contents are also beneficial for processing. It is assumed that harvesting takes place using a self propelled chopper, as this has been identified as the cheapest option in other studies (Smeets et al. 2009b). In addition, chopped Miscanthus dries more easily and this improves future processing. Because ethanol production is assumed for the entire year, a continuous biomass supply is required. Therefore, an average storage time of 6 months is assumed, with an average dry matter loss of 2% over this 6 months (Smeets et al. 2009b). The ‘chips’ are assumed to be transported to a lignocellulose ethanol plant by truck. After physical size reduction, the cellulose is broken down into free glucose molecules by enzymatic hydrolysis (Hamelinck and Faaij 2006). In the fermentation step, the free sugars are converted to ethanol. Within the timeframe considered by this study, dilute acid pre-treatment, on-site enzyme production, enzymatic cellulose hydrolysis, and a Simultaneous Saccharification and Fermentation (SSF) configuration boiler and steam turbine, are expected to be the most prominent technologies for converting lignocellose crops to ethanol (Hamelinck et al. 2005a). The characteristics of this conversion pathway are therefore assumed for this study. From an economic perspective, large scale facilities are preferable to small scale facilities and a capacity of 400 MW is generally required to achieve reasonable production costs (Hamelinck et al. 2005a). Therefore, it is assumed that Miscanthus is processed in a 400 MW ethanol plant (640 kton oven dry ton annual input, 4 PJethanol) that is located close to the port of Eemshaven. Since the expected regional feedstock supply does not meet the input requirements of a plant of this size, it is assumed that 70% of the lignocellulosic material will have to come from international supply chains. It is assumed that the import of lignocellulose material from abroad is feasible because of relatively low production costs, because pre-treatment can be applied, and because of the relatively low cost of international sea transport (Hamelinck et al. 2005b). In order to put these two bioenergy chains into context, other ethanol production chains are also assessed. Ethanol from currently cultivated annual crops such as wheat and maize is considered as well is ethanol from perennial crops such as switchgrass and willow. Specific data regarding required field operations, seed fertilizer and pesticide application, yield levels and dry matter content, transport of biomass and conversion to ethanol are provided in the Appendix.

2.3 Method The competitiveness of the bioenergy chains is assessed by comparing the economic performance of the bioenergy crops with the current use of agricultural land and by 40

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

comparing the production cost of bioethanol with average petrol prices. The way the spatial distribution of soil suitability and current agricultural land use affect competitiveness of bioethanol chains is addressed as follows. First, calculation methods for the economic competiveness of crops and the production costs are discussed. Thereafter, the method to determine the spatial distribution of soil suitability for individual crops and the effect on economic performance is considered.

2.3.1 NPV calculations for crop production In order to compare both annual and perennial crops, all costs and benefits during the cultivation phase are discounted and aggregated to provide their Net Present Value (NPV) (see Equation 2.1). The NPVs of the various rotations are calculated by multiplying the NPV of the individual crops by their proportional share in the rotation (see Table 2.1). N

∑(I

M

⋅ B ) − ∑ ( Jmy ⋅ C m )

ny n Y =x = n 1= m 1 cr y Y =1

NPV = ∑

(1 + a )

Equation 2.1 NPVcr I B J C a y

Net Present Value of crop per ha Occurrence of positive monetary flow n in year y Revenues of monetary flow n per ha Occurrence of negative monetary flow m in year y Cost of monetary flow m per ha Discount rate Annuity period

€/ha # €/ha # €/ha % y

The annuity time period considered here is 20 years, which is in line with the lifetime of the perennial crops and the lifetime of conversion plants (see Table 2.13 of the Appendix). A discount rate of 5.5% is assumed. This is a realistic interest rate for farmer loans (de Wolf and van der Klooster 2006 and J. Houtsma pers. comm.); but is considered to be low for commercial investment projects. For pasture, the NPV of grassland is compared to the NPV of Miscanthus. The revenue from pasture is represented by the avoided cost of fodder and the benefits related to manure application. For arable land, the best rotation for the specific soil type (clay/sand) is compared to a rotation with an increased proportion of sugar beet and to Miscanthus. The costs and revenues of crop production depend on soil and climate, the economic environment, and the farm management system. All these variables are regionally specific. For the calculation of the economic performance of crop production, only costs and 41

SHADES OF GREEN

benefits directly related to cultivation are taken into account. Overhead costs and general farm activities (e.g. maintenance of barns and farm area, cleaning, and administration) are not considered in this study. The costs related to crop production generally include four main categories of expenses: land costs, field operation costs (contractor, machinery, labour and diesel costs), input costs (seeds, fertilizers and pesticides) and fixed costs (insurance, soil sample assessment, etc). The benefits of crop production are the revenue from: selling the main product, selling the co-product, CAP subsidies for crop production. In our study, all costs and revenue are based on price levels for 2006 and are included in the supplementary on-line material. The lease price of land is used to reflect the land cost for farmers. A large variety of field operations needs to be carried out for the production of crops: soil preparation, seeding/planting, fertilization, weed and disease control, harvesting, storage, and drying. Machine costs for field operations are derived from de Wolf and van der Klooster (2006) and account for purchase price, salvage value, lifetime, interest rate, average annual operating hours, maintenance and repair, storage, insurance cost, and work rate of the specific field operation. The fuel use per field operation is related to the type of machine used for the operation and the work rate. The most commonly used tractor capacities for specific field activities are based on de Wolf and van der Klooster (2006). For field operations that are commonly outsourced in this region (e.g. seeding and harvesting beet and maize), contractor prices are incorporated. The contractor prices include costs for machinery, labour and fuel. For non-outsourced field activities, farmers’ labour costs are assumed for the first worker, while for every additional worker, labour costs for an average employee are assumed. The cost of harvesting perennial crops are related to the per hectare yield levels. The relationship between yield levels and harvest costs is non-linear and is described for willow by the wood supply research Group of the University of Aberdeen (WSRG 1994). It is assumed that this relationship also applies for other perennial crops (Smeets et al. 2009b). See Equation 2.2. HC 4.33 ⋅ Y −0,589 = Equation 2.2 HC Y

Harvest costs Yield

€/ha odt/ha

The fixed costs are a compilation of several costs that occur annually. These depend on the crop type, and include the costs for insurance, soil sample assessment, certifying and crop testing, tare, prevention of erosion, and national product levy. The input costs consist 42

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

of the cost for planting material, fertilizers, and pesticides and are determined by the application rates and costs per unit. The revenue for the farmer consists of the sale of products and CAP subsidies. For cereals, both main- and co-products have market value.

2.3.2 Cost of ethanol In order to calculate the ethanol production costs, all costs and benefits during all stages of the supply chain need to be taken into account. The specified cost calculation for perennial crops making use of the NPV has been demonstrated by van den Broek et al. (2000b). In general, only monetary flows can be discounted. However, since the yield represents a monetary flow, it is legitimate to discount this output too (van den Broek et al. 2000a). The allocation of feedstock production costs is based on the economic value of the main- and co-product (e.g. straw). All costs related to loading, unloading and transport need to be calculated per ton of product. This includes the cost of labour, fuel and depreciation of machinery. Finally, the costs and revenue for ethanol production need to be taken into account. This includes investment costs (depreciated over the lifetime), operations and maintenance (O&M) costs and the costs for fuel, gas, electricity and other inputs needed for the process. Benefits include revenue from co-products or electricity produced during processing. The scale, load factor and efficiency determine the annual input (feedstock) and output (ethanol). Equation 2.3 shows how the ethanol production costs are calculated.

C eth

 CC ⋅ a   + OM + EC − CP  y  ( Fs + Tr ⋅ D ) / ( Dm ⋅ E )  1 − (1 + a )  + η plant AOeth Equation 2.3

Ceth Fs Tr D Dm E ηplant CC a y OM EC CP AOeth

Cost of ethanol Feedstock costs Feedstock specific transport costs Distance to plant Dry matter content of feedstock LHVdm feedstock Plant efficiency (GJinput/GJoutput) Capital costs Discount rate Lifetime Annual O&M costs Annual energy input costs Annual revenues co-products Annual output ethanol

€/GJ €/ton (fresh) €/ton/km (fresh) Km % GJ/odt % € % Y €/y €/y €/y GJ/y 43

SHADES OF GREEN

2.3.3

NPV and costs of feedstock differentiated for soil suitability

Crop yields vary within the region due to different soil qualities. Therefore, the NPV of crops and the costs of feedstock are differentiated for different soil quality classes. To map the soil suitability and the related yield for the different crops in our assessment, we use the most recent HELP (Her-EvaluatieLandinrichtingsProject) system (Brouwer and Huinink 2002; Brouwer et al. 2003). In this method, physical yields are determined by a combination of soil characteristics (e.g. water holding capacity, clay-sand-peat contents, rooting depth and stoniness) and water tables in summer and winter. The total yield reduction (Dtot) relative to the maximum potential yield is determined by the yield reduction caused by drought (Ddr) (mostly in summer) and the yield reduction caused by water surplus (Dwa) (mostly in winter) assuming no irrigation. See Equation 2.4.

 100 − Dwa  Dtot = Dwa +  ⋅ Ddr   100  Equation 2.4

The yield level reductions were produced for the most common arable crops and mapped by (Brouwer and Huinink 2002) onto grid with 25 x 25m cells using GIS (Geographic Information System). In the present HELP system a large selection of crops is included, but perennial biomass crops are missing and so are seed potatoes, summer wheat, barley and rape seed (see Table 2.3). Estimates of yields losses of the missing crops ware made based on existing tables in combination with crop need knowledge. The expected yield loss of Miscanthus due to water and drought is based on Christian et al (2001) Lewandowski et al (2003) and W. Elbersen (pers. comm.). The assumptions regarding yield reductions due to water and draught stress of annual and perennial crops are summarised in Table 2.3. The crop specific HELP tables are used to map the soil suitability for all individual crops. This results in separate map layers of crop specific yield reductions. The suitability classes used are depicted in Table 2.4. The potentially suitable area includes the whole agricultural area excluding land used for greenhouses and land within Natura 2000 conservation areas. Yield statistics provided by LEI and CBS (2007) and de Wolf and van der Klooster (2006) present average yield levels for the region, differentiated to sand and clay soils. These average yield levels are translated to yield levels per suitability class by taking the yield reduction per suitability class and the relative share of suitability class per crop for current land use into account. 44

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

Table 2.3: Crops included in the HELP system (Her-Evaluatie van LandinrichtingsPlannen – Reevaluation of spatial planning) and new crops introduced including their relative sensitivity to drought and water damage.

Summer wheat Winter wheat

Included in HELP No Yes

Summer barley

No

Winter barley Feeding potatoes Seed potatoes

No Yes No

Industrial potatoes Sugar beet Rape seed Maize

Yes Yes No Yes

Miscanthus

No

Switchgrass

No

Willow

No

Grass

Yes

Perennials

Annuals

Crop

Assumed water and drought sensitivity The same as winter wheat, but more sensitive to drought. Derived from (Brouwer and Huinink 2002; Brouwer et al. 2003) The same specifications as winter wheat, but more sensitive to drought. The same specifications as winter wheat Derived from (Brouwer and Huinink 2002; Brouwer et al. 2003) More sensitive to both drought and water damage then feeding potato Derived from (Brouwer and Huinink 2002; Brouwer et al. 2003) Derived from (Brouwer and Huinink 2002; Brouwer et al. 2003) The same specifications as summer wheat Derived from (Brouwer and Huinink 2002; Brouwer et al. 2003) The same sensitivity to excess water as maize (Christian et al. 2001), but with slightly lower yield losses for dry conditions because of its deeper rooting system. It has a high tolerance to severe water stress conditions (Monti et al. 2008). Therefore it is expected to be more drought tolerant then Miscanthus (and certainly willow) and similarly tolerant to wet circumstances as Miscanthus. Willow can withstand seasonal flooding but not permanent water-logging (DEFRA 2002). It is expected to be more tolerant to wet circumstances and more sensitive to drought then Miscanthus and switchgrass. Derived from (Brouwer and Huinink 2002; Brouwer et al. 2003)

Table 2.4: Classification soil suitability as function of yield reduction due to water and drought stress. Suitability classification

Yield reduction

Very suitable

0-10%

High suitable

10-20%

Suitable

20-30%

Medium suitable

30-40%

Low suitable

40-60%

Marginally suitable

60-80%

Very marginally suitable

80-100%

45

SHADES OF GREEN

The management responses to yield reductions are not always clear. On the one hand, fertilizer inputs may be lower, due to reduced crop removal from the field during harvesting. On the other hand, the efficiency of fertilizer uptake may be decreased on poorer soils, resulting in increased application requirements. In the case of herbicides, applications may be higher on better soils, since weeds are likely to generate more biomass. However, since the crop canopies may close earlier in the growing season on better soils, the crop is better able to compete with weeds, which could reduce herbicide requirements. Because the management response to yield reductions can result either in an increase or a decrease of inputs, and because management is also dependent on local circumstances and individual decisions, no general rule regarding the level of input response to yield reductions can be made (A.J. Haverkort, J.G. Conijn and J.J. Schroder, pers. comm.). Therefore, we assume that input levels remain constant over soils of different quality and that the revenue achieved determines whether a crop is grown at a specific location. The NPV of the crops for each soil suitability class are linked to the crop specific soil suitability maps. For the NPV of rotations, individual map layers of the crops are combined for a final NPV map and weighted by the proportion of that crop in the rotation (Table 2.1). In addition to the NPV, the cost of feedstock production of Miscanthus and Sugar beet for every soil suitability class are linked to the GIS maps. All parameters used for the calculation of the competitiveness of bioenergy crops compared to current land use and the calculation of the cost of feedstock and ethanol production are provided in the supplementary on-line material.

2.4 Results The NPV of most agricultural crops, especially cereals, are found to be negative when all costs are included. In Figure 2.1 the proportion of costs and benefits (excluding subsidies) in the NPV of conventional crops, rotations, and perennial crops are shown for “very suitable” soils. Large differences are evident between intensively managed crops like potatoes and sugar beet, for which revenues are high but investments are high too, and less intensive crops like wheat and barley, which require far lower inputs and labour but do not provide high revenue.

46

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

Figure 2.1: Individual contributions of cost items and benefits to Net present Value (NPV) of individual crops, crop rotations, and perennial energy crops excluding subsidies.

In Figure 2.2, the NPV of perennial crops, typical rotations, and rotations with an increased proportion of sugar beet are shown for the different soil suitability classes. This figure shows that NPVs always decrease on less suitable soils and that the rate of decrease is greater for the crop rotations than for the perennials. This is due to the intensive management requirements of annual crops compared to perennial crops. Because it is assumed that inputs and work rates do not decrease for less suitable soils (except for yield related costs like harvest and drying), the economic performance of intensively managed crops declines more rapidly than the performance of less intensively managed crops on less suitable soils. Figure 2.2 shows that an increased share of sugar beet in rotations generally has no significant effect on the NPV, except for the ‘Clay II’ rotation. For this rotation, an increase in the proportion of sugar beet in the rotation causes a lower NPV on very suitable soils, but achieves a less negative NPV on less suitable soils, because sugar beet substitutes for potatoes, which have very high yield losses on less suitable soils. For less suitable soils (> 20% yield loss) the NPV of perennial crops exceeds the NPV of rotations. However, at this point the NPV of perennials is also low compared to keeping the land fallow. 47

SHADES OF GREEN

Figure 2.2: Net present value of perennials, typical rotation schemes and rotations schemes including an extra share of sugar beet (ES) for different soil suitability classes (excluding subsidies).

Currently, most farmers receive CAP (Common Agricultural Policy) support for cultivating agricultural crops (up to 446€/ha). Since energy crops receive little support (45 €/ha) in contrast to food and feed crops, the gap between the NPV of conventional crops and energy crops would increase on suitable soils. Thus, when subsidies would be included the intersection between perennial crops and conventional land use moves towards less suitable soils (> 30% yield loss). As noted previously, farmers often do not account for the cost of land, their own labour, and machinery. Omitting the cost of labour and machinery especially influences the NPV of intensively managed crops, and in these circumstances, perennials are only competitive on low and less suitable soils (> 50% yield reduction). Since perennial crops are more tolerant to water and drought stress, it is possible that some areas could be suitable for perennial crops and less suitable for rotation crops. The significance of this can only be depicted spatially. Therefore, the NPVs of all the crops (including all costs and excluding subsidies) were linked to the soil suitability maps for the individual crops. For the NPV of rotations, individual map layers of the crops were combined for a final NPV map and weighted by the proportion of that crop in the rotation. The mapped NPV of rotations on clay and sand, pasture and maize were then combined with a map of current land use and, for clay and sand, the best performing rotations were then selected. Then, over the whole agricultural area, the NPV of current land use was compared with the NPV of Miscanthus on a map of 25m x 25m cells (Figure 2.3). The same was done for the increased sugar beet rotation, but since there is little difference between the economic performance of extended sugar beet rotation and the conventional rotations (see Figure 2.2), this map is not presented here.

48

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

The green areas in Figure 2.3 indicate where Miscanthus can compete with current land use because of a higher NPV. Most of these areas are currently in use for pasture and are often too wet for arable crops. The red areas reflect those zones in which current land use is most profitable. These zones have fertile soils and are well suited for cultivation of profitable crops like potatoes and sugar beet. In these locations, it is very unlikely farmers will be willing to switch to energy cropping systems, at least from an economic perspective. Table 2.5 shows that cropping Miscanthus on land that is currently used for pasture is often more profitable than current practice, but that Miscanthus is almost never more profitable on land that is currently used for maize. Table 2.5 also shows that Miscanthus is more likely to be competitive with rotations on sandy soils than rotations on clayey soils.

Figure 2.3: Map of ΔNPV (= NPV of current land use - NPV perennial energy crops) for the whole agricultural area of the North of the Netherlands. Negative value (green areas) indicates where Miscanthus has a higher NPV than current land use. All cost items are included and subsidies are omitted.

49

SHADES OF GREEN

The distribution of those areas where Miscanthus is competitive with current land use is depicted as a function of the soil suitability classes in Table 2.6. Most of the land on which Miscanthus is competitive with current land use is moderately suitable for Miscanthus (90% of the total area is low to highly suitable for Miscanthus). This is plausible, since ‘very marginally suitable’ and ‘marginally suitable’ soils are very rare, and ‘very suitable’ soils are often more suitable for conventional cropping systems, which achieve higher NPVs on these soils.

0.00

0.15

Rotation sand Maize Grass Total proportion of land of highest NPV

Miscanthus

0.58

Grass

0.85

Maize

0.00

Rotation Sand II

Rotation Clay II

Rotation clay

Rotation Sand I

Current land use

Rotation Clay I

Table 2.5: The proportion of land that is more profitable under Miscanthus or more profitable under the current land uses of arable crop rotations on clayey soils and sandy soils, maize and grass.

0.42 0.97

0.00

0.15

0.17

0.00

0.03

0.01

0.12

0.88

0.06

0.61

Very marginally suitable

Marginally suitable

Low suitable

Moderately suitable

Suitable

High suitable

Very suitable

NPV Miscanthus > NPV current land use

Table 2.6: Share of area where Miscanthus has higher Net Present Value than current land use (ΔNPV is negative) in total and for different suitability classes.

Miscanthus-clay rotation

0.00

0.00

0.02

0.01

0.01

0.00

0.00

0.04

Miscanthus-sand rotation

0.00

0.00

0.05

0.02

0.05

0.08

0.00

0.20

Miscanthus-maize

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

Miscanthus on land currently in use for:

Miscanthus-pastures 0.01 0.01 0.20 0.08 0.19 0.19 0.08 0.76 Miscanthus Total 0.01 0.01 0.26 0.11 0.25 0.27 0.09 1a a Total share of land where Miscanthus has higher Net Present Value (NPV) than current land use corresponds with 0.61 of total agricultural area (see total Table 2.5).

50

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

2.4.1 Cost of biomass The cost of biomass is expressed per GJ feedstock (Lower Heating Value of DM of the whole crop) at the farm gate, and is differentiated for each crop and soil suitability class (Figure 2.4, left). In the Northern region of the Netherlands, Miscanthus has a total potential energy yield of 155 PJ, if the whole agricultural area is dedicated to this crop. The lowest production cost is 5.4 €/GJ on very suitable soils, the highest is 41.6 €/GJ on very marginally suitable soils. The potential energy yield from sugar beet for the whole agricultural land area (134 PJ) is smaller than for Miscanthus. Also, the cost of production (9.7 €/GJ and above) is higher than for Miscanthus, but lower than for most other annual crops. However, if it is assumed that sugar beet is only cultivated on land currently in use as arable land, as assumed in the Refuel study (de Wit and Faaij 2010), the potential would greatly decrease. In addition, only a maximum share of 25% in the rotation is permissible, which would decrease the potential even more. An additional issue is that biomass production costs in the North of the Netherlands are likely to greatly exceed the cost of biomass imported from abroad (Lewandowski and Faaij 2006), which in the case of lignocellulosic biomass, are expected to vary from 3.0 - 3.5 €/GJ for pellets from Latin America, 3.5 - 5.0 €/GJ for pellets from Eastern Europe, and 4.5 - 6.5 €/GJ for pellets from Scandinavia (Hamelinck et al. 2005b).

Figure 2.4: Cost supply curves for various crops in the North of the Netherlands for the total of agricultural land in the region. The first ‘step’ in the curves indicates the cost of biomass produced on very suitable soils, the second for high suitable… the last step of each curve indicates the cost of biomass produced on very marginally soils (left). Cost supply curve of Miscanthus based on land availability from ΔNPV (Net present value) and distribution over soil suitability and the potential related to the land availability according to the Refuel study (right).

Taking into account the figures presented in Table 2.5 and Table 2.6, a cost supply curve can be constructed for Miscanthus for the area where its cultivation is competitive with current land use (Figure 2.4, right). The potential presented in the figure is relatively high. The data presented in the Refuel study regarding land availability for bioenergy crops (see 51

SHADES OF GREEN

Section 2.2.2) indicate that only a small part of this potential can be exploited for bioenergy crops, without diminishing the self-sufficiency of the region (the ‘optimistic’ scenario of Refuel is presented by black dots in Figure 2.4, right). Assuming that only the least cost production areas are likely to be dedicated to bioenergy crops (Figure 2.4, right), this results in a potential supply of 2.7 PJ at a cost of 5.4 to 5.9 €/GJ compared with a potential of 71 PJ at a cost of 5.4 to 9.4 €/GJ if all the land where Miscanthus is competitive with current land use is taken into account. A cost supply curve for bioenergy feedstock from sugar beet for the area where it is competitive with current land use cannot be made, as rotations that exceed the sugar beet quota have lower returns than current rotations. The cost of feedstock production is affected by the soil suitability. In Figure 2.5, the spatial distribution of the cost of sugar beet and Miscanthus production are given. Both crops achieve lowest production costs in the Northern area of the region. A relatively large area achieves comparatively low production costs for Miscanthus. The production costs of sugar beet are generally higher and increase more rapidly in less suitable conditions. In Figure 2.5 (top), the potential cost of sugar beet cropping on land now used for pastures is depicted as well. However, these areas are considered to be unavailable for sugar beet production. In this Figure, the land currently used for pasture is mainly coloured dark (very high production costs). Comparing Figure 2.3 with Figure 2.5 (bottom) shows that for some locations where Miscanthus performs better than current land use, production costs are very high. However, most areas where Miscanthus has a higher NPV than current land use have relatively low production costs. These are the most promising locations for Miscanthus production.

2.4.2 Costs of ethanol In Figure 2.6, the cost of ethanol production (€/GJ) from sugar beet and Miscanthus are represented. This figure is based on the least cost feedstock produced on very suitable soils (all costs factors including land, labour and machinery are taken into account). For comparison, the cost of ethanol from wheat and maize and the cost of petrol are depicted as well. The petrol prices do not include VAT, excise and margins. The difference between the cost of bioethanol and petrol is significant (>182%) assuming an oil price level of 62 US$/barrel. However, when oil price levels increase to 100 US$/barrel (average level of 2008) (OECD and IEA 2008) or 150 US$/barrel (as projected for 2020 by OECD and IEA (2008)), bioethanol could become competitive to petrol. The bioenergy feedstock costs vary for the various crops produced at suitable soils (range 5 - 10 €/GJ, see Figure 2.4). Conversely, the cost range of ethanol production is relatively small (24 - 27 €/GJ, see 52

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

Figure 2.6). This is caused mainly by the fact that relatively expensive feedstock in the form of sugar and starch crops require less advanced technology for the conversion to ethanol. The contribution of capital and operations and maintenance (O&M) cost are relatively large for ethanol production from lignocellulosic crops. The distance from field to processing plant is assumed to be the same for all feedstock, but the share of transport costs for ethanol from sugar beet is large due to the high moisture content of sugar beet. For production of ethanol from wheat, only the main product (grain) is used; straw is considered to be a co-product used for other purposes. Currently there is a relatively high demand for straw for several purposes (stables, crop coverage, etc). Allocation of costs for production of the ‘main products’ is based on economic value. Based on this allocation, the production cost of straw exceeds the cost of Miscanthus (€/GJfeedtock). Therefore the cost of ethanol production from straw will be higher than from Miscanthus and will not be profitable. For this reason, it is assumed that demand for straw for ethanol production is not yet an additional competitive factor in the market for straw. The leaves and crowns of sugar beet are assumed to be left on the field, and are therefore not considered to be coproducts. For lignocellulose crops, the whole crop is used for conversion to ethanol. If the total area where Miscanthus is competitive with current land use is dedicated to Miscanthus for ethanol, 25 PJ ethanol could be produced annually. However, the Refuel study indicates that only a minor share can be used for bioenergy crops before compromising self-sufficiency. This results in an annual production of 1 PJ of ethanol at a cost between 24.4 and 25.9 €/GJ, equivalent to 0.7 % of the energy in the petrol used in the Netherlands (142 PJ) in 2006 (CBS, 2008).

53

SHADES OF GREEN

Figure 2.5: Spatial distribution of sugar beet production costs (top) and Miscanthus production costs in €/GJ.

54

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

Figure 2.6: Cost of ethanol production from various feedstock in the North of the Netherlands compared to petrol prices for various oil price levels (US$/Barrel). Least cost feedstock produced on very suitable soils are incorporated. (CAPEX: capital expenditures, O&M: Operations and Maintenance costs).

2.4.3 Sensitivity analysis In this section the sensitivity of the NPV, the cost of biomass and the cost of ethanol for various key parameters is assessed. These have been selected because of expected fluctuations or uncertainty in specific parameters (e.g. commodity prices, fuel prices and discount rate) and/or the expected effect of the key parameter on the final result (e.g. biomass yield and labour wages). In Figure 2.7, the sensitivity of the NPV of Miscanthus and sugar beet cultivated on very suitable soils is presented using spider diagrams. The NPV of Miscanthus and sugar beet are very sensitive to changes in yield levels and market prices. The NPV of sugar beet is more sensitive to changes in labour and energy prices than Miscanthus due to the relatively intensive management that is required. Biomass costs are sensitive to changes in yield, especially in the case of lower yields, where costs increase significantly. Miscanthus production cost is sensitive to changes in the discount rate. This is due to the high initial investment required and to the relatively long period of time that it takes to achieve high yields. For sugar beet the discount rate has little effect, since costs and benefits are approximately equal every year. The cost of ethanol production is very susceptible to yield levels and efficiency. The impact of higher energy prices is different for the cost of ethanol production from Miscanthus and from sugar beet. When energy costs increase, the costs of ethanol production from sugar beet also increases, due to higher feedstock and transport costs. For Miscanthus these costs also increase, but the co-product of ethanol production of Miscanthus, electricity, 55

SHADES OF GREEN

increases in value too. Therefore, for lignocellulosic ethanol, the net effect is a decrease in ethanol production costs when energy prices increase. The sensitivity for yield level represents the sensitivity for changes in soil suitability.

Figure 2.7: Sensitivity analysis for Net Present Value (NPV), cost of biomass, and cost of ethanol of Miscanthus and sugar beet. Key parameters, discount rate, energy prices, labour wages, yield levels, commodity prices and efficiency of conversion, are varied between 50% and 200% of the original value.

56

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

2.5 Discussion 2.5.1 Method and input data In this study it is assumed that economic performance is the main driver for the adoption of different agricultural crops by farmers. The personal preferences of farmers, which can also influence land use, are not included in our study. Other factors that can influence the economic performance of land use, such as previous investment in crop specific machinery and equipment, long-term agreements with procurers of processing chains, individual management and rotation choices, and additional costs or benefits of specific land use due to locally enforced policy measures and subsidies (e.g. to protect ecosystems and historic landscapes), are also not included in this study. An important assumption is that here inputs of seeds, fertilizers, pesticides, and field operations do not change for different soil suitability classes. The main reason for this is that poorer soils can require both higher and lower levels of inputs and based on available data, no general trend can be distinguished. The contribution of input costs to the total feedstock costs is relatively low for perennial crops (about 6%) but more significant for annual crops (about 12% for barley and 30% for feeding potatoes). If it were to be assumed that fewer inputs would be applied to crops on less suitable soils, the feedstock costs would decrease for these poorer soils. A further issue lies in the scale of analysis used in this study, which here is based on a one-hectare comparison of different crops. Farmers, however, need to consider the whole farm business and the way in which individual enterprises link with each other. For example, this could have implications for the analysis of the pasture areas, since only the replacement value of fodder and the application of manure are considered as economic benefits, and other benefits, such as subsidies and revenue from cattle breeding have not been included. Although the NPV does not necessarily represent every individual farmer’s perspective, it does present a broad economic picture of the relative profitability of different land uses and, as a result, provides an indication of how land use might change at a regional level. Therefore, we propose that those areas where bioenergy crop production has been found to be relatively profitable in this study, could serve as a starting point for economic analysis of bioenergy production at a farm level. There is little experience with the cultivation of perennial bioenergy crops in the Netherlands and as a result, management practices have seen little development or optimisation in comparison with conventional crops, where management has been optimized over the decades. For perennials, there are uncertainties regarding input requirements (e.g. rhizomes and fertilizer needs) and attainable yield levels, which have

57

SHADES OF GREEN

large implications for economic performance. This uncertainty is also reflected by the large differences for input requirements in the literature. In addition, since ethanol plants based on lignocellulose feedstock are not commercially running yet, efficiency and investment costs used in this study come with some uncertainty attached.

2.5.2 Results The NPV of crops are very sensitive to market prices of agricultural products. These prices have fluctuated to a large extent over the last few years. The FAO food price index increased from 116 to 219 between 2006 and spring 2008 and then decreased to 148 in December 2008 (FAO 2009b). Therefore, the results related to the prices used here need to be carefully interpreted. Our assessment indicates that Miscanthus could be competitive with current land use in a relatively large area (given a level playing field in terms of subsidies). The maps show that the area where Miscanthus could be competitive with current land use is dominated by pastures. However, since there are uncertainties regarding management data of pastures and additional benefits, and differences in NPVs are small; this result should be interpreted with care. The Refuel study also indicates a marginal availability of land currently in use for pastures (de Wit and Faaij 2010). Therefore, the actual area where bioenergy crops are competitive with pasture is expected to be very limited. The maps in Figure 2.3 and Figure 2.5 give an indication of which areas could become the most promising areas for energy crop production. These areas are likely to be the ones where the NPV of Miscanthus is higher than the NPV of current land use and the costs of feedstock production are low. The European sugar market is protected by the European Union by a set quota and a guaranteed intervention price. Intervention prices of white sugar were reduced from 63 €/100kg sugar in 2006 to 42 €/100kg sugar in 2009 (Berkhout and van Berkum 2005). The economic value of 1 ton of sugar beet for sugar therefore decreased from 82 €/ton in 2006 to 55 €/ton in 2009 compared with an economic value of 53 €/ton for ethanol (assuming an ethanol price of 0.60 €/l). This shows that the production of sugar has become less profitable over the years, as a result of the reduced intervention prices for sugar. In addition, when the EU market opens to imports from abroad, ethanol production (and other uses of sugar beet) could become more attractive. For example, sugar beet can be used for (potential) applications in food, feed and the biochemistry industry. More advanced products (e.g. amino acids) with higher market value could be produced from sugar beet in combination with ethanol. This could also contribute to a larger greenhouse gas and fossil fuel mitigation potential (Brehmer et al. 2009).

58

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

Production costs of bioethanol from Miscanthus are relatively high (24 €/GJ) compared to current petrol prices (12 €/GJ). Feedstock production costs of domestic cultivated Miscanthus would need to be reduced by 38% to 3.33 €/GJ to be able to achieve ethanol production costs that could compete with petrol prices (oil price 62$/barrel). The ethanol production costs from Miscanthus in the Netherlands are equivalent to the prices of ethanol imported from Brazil, mainly due to a high import duty of almost 5 €/GJ. With improvement in technology and management, ethanol production costs could be reduced to about 13.5 €/GJ in the future (Hamelinck et al. 2005a; Hamelinck and Hoogwijk 2007; de Wit and Faaij 2010). In addition, according to the World Energy Outlook, oil prices are likely to increase (OECD and IEA 2008). Therefore, bioethanol is expected to become more competitive with petrol in the future. In this study, we have compared the economic performance between current land use and bioenergy crops. Although the influence of subsidies has been assessed, the main comparison is based on cost calculations that exclude subsidies. It should be noted, however, that the current land use is a result of (historical and current) agricultural policies and subsidies. In order to achieve the feedstock production cost of 3.33 €/GJ for Miscanthus (at which ethanol production could compete with petrol prices), a subsidy of 600 €/ha is required. At that subsidy level, Miscanthus is more profitable than pasture and all crop rotations (including subsidies) on every soil suitability class, except for clay rotations on very high and high suitability soils. For sugar beet, a subsidy of 1080 €/ha is required to achieve a feedstock production cost (5.68 €/GJ) at which ethanol production costs could compete with petrol prices. At this subsidy level it is economically attractive to increase the share of sugar beet in all rotations for all soil suitability classes. The potential contribution from domestically produced ethanol from Miscanthus and/or sugar beet is relatively small (<1% of total energy use in the transport sector) assuming that only the ‘available land’ as indicated by the Refuel study can be used for bioenergy crops. Therefore, the Netherlands will have to rely on imported biomass/bioenergy to meet its targets for biofuel use in transport (10% in 2020) and renewable energy (20% in 2020) (Projectgroep 'Duurzame productie van Biomassa' 2006).

2.6 Conclusions In this paper, the potential and economic viability of bioethanol chains in the Northern region of the Netherlands has been assessed for different soil suitability classes. The results have been compared to current agricultural land use. In addition, the spatial distribution of feedstock production and the production costs have been mapped. With

59

SHADES OF GREEN

this approach, we have assessed where land use changes in favour of bioenergy crops are most likely to occur. The results of the NPV calculations show that an increased share of sugar beet for ethanol production cannot compete with current cropping systems under present quota conditions and commodity prices. The potential biomass production from sugar beet is lower than from Miscanthus, since only arable land is assumed to be appropriate and less well suitable land is available for sugar beet cultivation. Most cost effective sugar beet production is on very suitable soils in the coastal area in the North and the East of the region. Ethanol from domestic produced sugar beet is significantly more expensive than petrol or ethanol produced from feedstock imported from abroad. Therefore, there are no economic incentives to produce sugar beet for ethanol production in the North of the Netherlands under current circumstances. However, when oil prices increase and ethanol production is combined with the production of more advanced products (e.g. bulk chemicals), the competitiveness could increase. The spatial analysis shows a large area in the North of the Netherlands where cultivation of Miscanthus could compete with current land use when a level playing field is established (i.e. when subsides are excluded). Ethanol production of Miscanthus appeared to be the least cost option of bioethanol production of domestically cultivated feedstock in this region, but is still almost twice as expensive (24.4 €/GJ ethanol) than petrol (12.3 €/GJ, at an oil price level of 62 US$/barrel) or ethanol produced from feedstock imported from abroad. Therefore, there are no economic incentives for the production of Miscanthus is the North of the Netherlands for ethanol production under current circumstances. However, if bioethanol production costs decrease because of technological learning and crude oil prices increase, bioethanol could become competitive. Taking the land availability of the Refuel study into account, the contribution of ethanol from domestic cultivated feedstock would be less than 1% of the petrol use in the Dutch transport sector. This indicates a marginal potential for biofuel chains in this particular region, but this can still contribute to meeting the fuel blending targets in the Netherlands for the near future. In the analysis of the competitiveness of Miscanthus production with current land use, current pasture land appeared to be an important potential area for Miscanthus cultivation. However, as indicated in the discussion, there are uncertainties regarding the economic performance of pastures at a farm level and additional research is required. Also a more in depth assessment regarding the relation between management, soil suitability and yield levels is needed in order to draw firmer conclusions concerning the economic 60

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

and practical viability of cultivation of bioenergy crops in the identified promising areas. Since combined production of advanced products and ethanol from biomass feedstock could be more beneficial than ethanol production alone in terms of economic performance and greenhouse gas mitigation potential, innovative biomass supply chains could be an interesting topic for further research. This study provides a generic methodology to identify promising locations for bioenergy crop production based on soil properties and current land use. The method can therefore be applied in other geographical regions and at higher levels of analysis. The most important conclusion from this assessment is that the spatial variation of economic viability of bioethanol production chains indicates where land use changes are most likely to occur. However, economic performance is just one of the criteria needed to investigate the sustainability of bioenergy production. The environmental impacts in relation to the spatial characteristics of regional bioenergy chains are also very important and need further investigation.

2.7 Acknowledgements This study is part of the Climate changes spatial planning program and is funded by the Dutch government, the European commission and Shell. The authors gratefully acknowledge the contributions to data collection of management and inputs of crop systems, pastures and perennial crops of Remco Schreuder, PPO Lelystad, Wageningen University and Research centre; Aart Evers and Peter Roelofs of Animal Science Group Lelystad of Wageningen university and research centre; Anton Haverkort, Jaap Schroder and Sjaak Conijn of PRI Agrosystems research of Wageningen university and research centre; Edward Smeets of the Copernicus institute, Utrecht University. In addition, the authors sincerely thank the contribution to the construction of the GIS maps of Michiel van Eupen and Rob Schmidt of Alterra.

61

SHADES OF GREEN

2.8 Appendix: Input data for calculation of economic performance of bioenergy crops 2.8.1 Cultivation In Table 2.7 and Table 2.8, the field operations required for annual crops and perennial crops are presented. The data on field operations and inputs for annual crops are based on (Schreuder et al. 2008) and the related cost figures are based on (de Wolf and van der Klooster 2006). The field operations and inputs for pastures and the related costs are based on data from the (Animal science Group 2005). Data on the field operations and inputs of Miscanthus, switchgrass and willow are from a variety of sources, including (Huisman et al. 1997; Bullard 2001; Bullard and Matcalfe 2001; Christian et al. 2001; DEFRA 2002; Heller et al. 2003; Lewandowski and Heinz 2003; Styles and Jones 2007b; Boehmel et al. 2008; Smeets et al. 2009b). The related costs of field operations and inputs for perennials are adapted for the Dutch situation using data from actual, though limited, experience in the Netherlands, and quantitative agricultural data of the Netherlands (de Wolf and van der Klooster 2006). The fuel used by agricultural machinery for field operations is assumed to be 3.67 litres per kWh on average (de Wolf and van der Klooster 2006). For a four wheel drive tractor, an average load factor of 0.8 is assumed, whilst for a self-propelled machine, a load factor of 0.9 is assumed. This results in a slightly lower rate of diesel use (+/-15%) than the approach proposed by the American Society of Agricultural Engineers (AAEA 2000), which is used by (Heller et al. 2003; Smeets et al. 2009b) (i.e. Diesel usage (l/h) = 0.22*Capacity [kW]). The cost of lubrication is assumed to be 10% of the fuel cost (de Wolf and van der Klooster 2006) and diesel used for agricultural purposes is exempted from taxes. Since little of the agricultural area in this region is irrigated (0-1% in wet years, 7-11% in dry years, primarily for potatoes and pastures) (Hoogeveen et al. 2003), irrigation costs are not included in this study. Although in practice, straw from cereals is sometimes left on the field and ploughed back into the soil, in this study it is assumed that straw is removed, since it has significant economic value (50-60 €/ton (de Wolf and van der Klooster 2006)). However, stubbles are left on the field and ploughed under. Since crowns and leaves of sugar beet are commonly left on the field and have no market value, in this study it is assumed they are not removed. It is assumed that nutrient requirements are met by fertilizer application and that no lime (Ca) or magnesium (Mg) is needed. In the first year of perennial crops (except for 62

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

pastures), no fertilizers are applied in order to minimize the growth of weeds (DEFRA 2002; Khanna et al. 2008). In Table 2.9 and Table 2.10 the input levels, average yields and price levels are depicted for both annual and perennial crops. Additional fixed costs are related to the yield levels and vary between 1% (cereals) and 8 % (sugar beet) of the total annual costs (de Wolf and van der Klooster 2006). The agricultural sector in the Netherlands is highly subsidized. Before 2005, farmers received crop specific subsidies linked to the amount of production. Since 2005, this system has been replaced by a Single Farm Payment reflecting historic subsidy payments provided for crops in the reference period between 2000-2002 (de Wolf and van der Klooster 2006). Table 2.7: Field operation for annual crops based on De Wolf and van der Klooster (2006) and Schreuder et al. (2008).

(c) Indicates whether the specific field operation is assumed to be performed by a contractor.

63

SHADES OF GREEN

Table 2.8: Field operations for perennial crops. Field operations Miscanthusa switchgrassa willowb grassc Soil Plough (1.2m) 1/20 1/20 1/20 1/20 Power harrow 3m 1/20 1/20 1/20 preparationd Cultivator 3m 1/20 Seeding and Rhizome planter 1/20 planting Precision drill 3m 1/20 4-row willow planter 1/20 Grass seed combination 1/20 Roller 3m 1/20 1/20 1/20 1/20 Fertilizer spreader >18me 19/20 19/20 5/20 60/20 Crop Organic manure spreaderf 40/20 nurturing Sprayer 24mg 10/20 10/20 7/20 23/20 Weed cultivatorh 1/20 10/20 Row cultivatori 1/20 j Harvesting Self propelled chopper 19/20 5/20 Adapted mower (coppice) 1/20 19/20 1/20 Baler 1/20 19/20 Circular grass mower 40/20 Circular rake 40/20 Grass yield turner 80/20 Soil recoveryk Rotary cultivator 1.2m 1/20 Sub soiler 1/20 1/20 1/20 Cultivator 2m 1/20 1/20 1/20 1/20 Spraying 24m 2/20 2/20 2/20 2/20 Bush hogger 1/20 Storagel 19/20 19/20 5/20 Storage in Silom 1 Storage a Field operations of Miscanthus and switchgrass are based on (Christian et al. 2001; Elbersen 2008; Smeets et al. 2009b) related costs are based on (de Wolf and van der Klooster 2006) and (Schreuder et al. 2008). b Field operations of willow are derived from (Venturi et al. 1999; DEFRA 2002; Heller et al. 2003; Heller et al. 2004; Styles and Jones 2007b; Boehmel et al. 2008) related costs are based on (de Wolf and van der Klooster 2006) and (Schreuder et al. 2008). c Field operations and related costs of grass are based on (Animal science Group 2005) and (Evers 2008; Roelofs 2008) For grass a reseeding rate of 10% is assumed (Animal science Group 2005). d For Miscanthus, switchgrass and willow only first year before planting. For grass 10% need to be ploughed and cultivated every year for reseeding. e Miscanthus and switchgrass every year (except first year). Willow every year after harvest. Grass every year. f Although manure could be applied on perennial energy crops, in this study it is assumed mineral requirements of energy crops are met by fertilizer application. For pastures, manure application and additional fertilizers is current practice. g Some studies indicate no chemical weed control is needed (Styles et al.), others indicate chemical weed suppressor is needed annually (Bullard and Matcalfe 2001). Miscanthus, switchgrass and willow are poor competitors in the first year after planting and therefore it is assumed that herbicides need to be applied in this period. In addition herbicides are applied every year after harvest (only in 50% of the cases for Miscanthus and switchgrass). The final year glyphosphate is applied in order to end the rotation. For grass chemical weed control is needed annually.

64

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

h

Miscanthus only in first year (). For willow 2 times in year of establishment and 2 times after each harvest (Venturi et al. 1999). i Willow only one time during year of establishment (Venturi et al. 1999). j Miscanthus and willow are coppiced by an adapted mower first year. After the cut back, Miscanthus is harvested every year, willow every fourth year by a self propelled chopper. Switchgrass is harvested every year from second year on with a mower. For grass it is assumed half of the yield is directly grazed by cattle and the other half is harvested in two cuts. After every cut one time raking and two times turning is needed. k At the end of the lifetime, the plants are destructed by spraying Glyphospate and by ploughing and cultivate the soil. For willow, after spraying and sub soiling the stools can be mulched into the top soil although this is a labour and machinery intensive process (DEFRA 2002). l It is assumed the chipped Miscanthus, switchgrass and willow are stored under a plastic cover (Smeets et al. 2009b) m The part of grass yield that is harvested, is stored in a silo for roughage.

65

Table 2.9: Inputs, yields and prices of annual crops. Source: (de Wolf and van der Klooster 2006)

SHADES OF GREEN

66

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

grassa

willow

switchgrass

Unit

Miscanthus

Table 2.10: Inputs, yields and prices of perennial crops.

Prices

Yields

Inputs

Seeds Quantity of seeds unit/ha 20000b 10c 1600d 10e Cost per unit €/unit 0.18f 2.70g 0.10h 35 j j k l Fertilizersi N kg/odt 2.2 5.9 6.9 210 P kg/odt 0.6 0.9 0.6 K kg/odt 6.5 2.9 2.0 m3/ha 40 Manure Sprays m 132n 149n 50o 65p First year €/ha/y Subsequent years €/ha/y 25 25 22.5 5.5 q r s t Average Yield Year 1 odt/ha 1.5 1.5 2.5 11.5 Year 2 odt/ha 7 6 10 11.5 Year 3 odt/ha 11 9 10 11.5 Year 4 odt/ha 14 11 10 11.5 Year 5 to 20 odt/ha 15 11 10 11.5 DM content (harvest) 0.70u 0.70u 0.5u 0.18v DM content (after field drying) 0.85 0.85 0.8 0.35 18w 18x 18y 16z LHV GJ/odt Market prices Main product €/odt 50aa 50aa 50aa 60ab Co-product 80ac a,d Subsidy €/ha/y 45 45 45 45ae a Assumptions for grass are based on (Animal science Group 2005; Evers 2008; Roelofs 2008) unless otherwise indicated. b Based on (Bullard and Matcalfe 2001; Styles and Jones 2007b; Monti et al. 2008; Smeets et al. 2009b) .In some studies a density of 10,000 rhizomes is assumed (Venturi et al. 1996; Himken et al. 1997; Huisman et al. 1997; Khanna et al. 2008; van der Voort et al. 2008). c Based on (Christian and Riche 1999; Bullard and Matcalfe 2001; Elbersen 2008; Smeets et al. 2009b) d Based on (Coeleman et al. 1996; Venturi et al. 1999). In some studies a density of 15 000 plantlets is assumed (Styles et al.; DEFRA 2002). e And 10% for reseeding for every successive year (Animal science Group 2005). f Based on (van der Voort et al. 2008; Smeets et al. 2009b). Although rhizome prices are expected to decrease to a level of 0,04 €/piece (Venturi et al. 1996). g Based on (Christian and Riche 1999; Smeets et al. 2009b). h Based on (Coeleman et al. 1996; Kuiper 2003). i In the first year of rotation of perennial crops, no fertilizers are applied in order to minimize the growth of weeds (DEFRA 2002; Khanna et al. 2008) except grass. High cumulative phosphate concentrations in the soil in agricultural areas in the Netherlands (Romkes and Oenema 2004) argue for no phosphate application for perennial crops (Elbersen 2008).

67

SHADES OF GREEN j

In [kg per odt harvested](Smeets et al. 2009b). Every year, the removed mass is replaced. Nutrient contents are based on delayed harvest, uptake efficiency accounted for is 0.8 for N, and 1.0 for P and K (Biewinga and Bijl 1996) (deposition of 17 kg/ha is accounted for). k In [kg per odt] (Biewinga and Bijl 1996) applied every year after harvest (year 2, 5, 9, 13, 17). Uptake efficiency of 0.8 for N and 1.0 for P and K is accounted for. l N fertilizer in kg/ ha. For pastures it is assumed that fertilization requirements are partly met by organic manure. Application of 40 m3 manure meets annual phosphate and potassium requirements and partly nitrogen needs. Concentration of 2.2 kg mineral N/m3 (uptake efficiency 95%) and 2,2 kg organic N/m3 (uptake efficiency 35%) is assumed (Animal science Group 2005; Animal Science Group 2008). m Miscanthus, switchgrass and willow are poor competitors in the first year after planting and therefore herbicides need to be applied in this period. Some studies indicate no chemical weed control is needed (Styles et al.), others indicate a annual chemical weed suppressor is needed (Bullard and Matcalfe 2001). In this study it is assumed that chemical weeding is needed in the first years and in 50% of the cases during rest of lifetime (Elbersen 2008). n Based on (Christian and Riche 1999; Bullard and Matcalfe 2001; Elbersen 2008; Smeets et al. 2009b) o Application rate in year 1(Heller et al. 2003). Subsequent applications: every year after harvesting (once every four years). p Application rate for year 1. Successive years 5 €/ha and every 5 years an additional 42 €/h (Animal science Group 2005). q (Christian and Riche 1999; Clifton-Brown et al. 2004; Elbersen et al. 2005). r (Christian and Riche 1999; Elbersen et al. 2005) first year no harvest takes place. s (WSRG 1995; Coeleman et al. 1996; Venturi et al. 1999). Harvested every four years. t (Duinkerken 2007) assumed is that half of yield is directly consumed by cattle and the other half is harvested (in two cuts) and stored in silo. u (Biewinga and Bijl 1996) (Venturi et al. 1999). Switchgrass is assumed to have the same dry matter content at harvest as Miscanthus. v Dry matter content after silage. w (Christian et al. 2001) x (Boehmel et al. 2008) y (Kuiper 2003) z (Kaltschmitt and Reinhardt 1997) aa Based on market prices of straw (de Wolf and van der Klooster 2006) ab Market price of fresh grass for roughage ac Market price of silo grass roughage ad (European Union 2007) ae There is no EU support for pastures, but there are several nature, biodiversity and bird protection policies for pastures. In addition, several support regulations for cattle are in place. However, these policies are farm and location specific. In case grass is cultivated for energy purposes, EU support for energy crops is applicable (Animal science Group 2005).

68

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

2.8.2 Transport The key data and main literature sources used to calculate the cost of feedstock transport are depicted in Table 2.11. The maximum capacity per truck is either limited by weight or by volume depending on the density of the transported biomass. Table 2.11: Parameters of feedstock transportation used in this study. Transport parameter Max load volume Max load weight Labour (un-)loading Capital costs unloading Fuel use loading Average speed Fixed cost per km Fuel use (77% loaded) Fuel use (empty)

Unit m3 t h/ton €/ton l/ton km/hour €/km l/km l/km

130 40 0.06 1.30 0.63 48 0,38 0.36 0.20

Reference (Hamelinck et al. 2005b) (Hamelinck et al. 2005b) (Smeets et al. 2009b) (Smeets et al. 2009b) (Smeets et al. 2009b) (NEA 2004) derived from (NEA 2004) (NEA 2004) (Smeets et al. 2009b)

2.8.3 Conversion In Table 2.12, characteristics of the conversion plants are depicted. The scale of the conversion plant is related to the expected supply derived from the Refuel study (see section 2.2.3) Table 2.12: Data about ethanol plants used in this study (Elsayed et al. 2003; Hamelinck 2004; Hamelinck et al. 2005a; Hamelinck and Hoogwijk 2007). Sugar beet

Scale

GJinput(LHV) /y

2995

1550

2075

11520

Scale factorc

R

0.67

0.67

0.67

0.7

Investment costs

M€

37.9

26.2

22.0

290

O&M

% investment

6,2

2.5

2.5

6.4

Load

h/y

8000

8000

8000

8000

Conversion efficiency

GJout/GJin

0.56

0.53

0.57

0.35

Electricity use

kWh/GJoutput

5.71

2.71

6.81

0

Fuel oil use

GJ/GJoutput

0.30

0.08

0

0

Natural gas use Co-product

GJ/GJoutput

0

0.44

0.82

0

Animal feed

Animal feed

Animal feed

Electricityd

0.85

0.06

0.04

29.14

f

f

ton/GJoutput

e

a

Wheat

Maize

Lignocellulosea

Unit b

Price co-product €/ton 3.6 114 114 0.19 In this study, the characteristics of near term lignocellulose conversion technologies are assumed (derived from (Hamelinck et al. 2005a)) . Because of technological development, future plants are expected to require lower investments costs and achieve higher efficiencies (Hamelinck et al. 2005a).

69

SHADES OF GREEN b

c

For sugar beet wheat and maize, the scale is adjusted to maximum land availability according to the Refuel study (see section 0) For lignocellulose a minimum scale of 400 MWinput based on partly imported biomass is assumed (Hamelinck et al. 2005a). The scale factor indicates the relation between increases in plant scale and associated decreasing production costs, according to a scale law which can be written as Cost  Scale  (de Wit et al. 2010). Scale factors are R

t

Cost 0

t =   Scale0 

based on Hamelinck and Hoogwijk (2007) and de Wit et al. (2010). Electricity production during process in [kWh/GJoutput] and prices are in [€/kWh]. Based on a electricity production efficiency of 4.1 % (Hamelinck 2004). e Prices range from 3.5 to 6.0[€/ton]. The lower price based on (Animal science Group 2005) and higher price on (Hamelinck and Hoogwijk 2007) which is derived from (Elsayed et al. 2003). The lower price is incorporated in this study because it reflects better the price levels of 2006 in the Netherlands. f Prices range from 114-148. The lower price of is based on the price of corn gluten feed (Animal science Group 2005) which is the most suitable equivalent of DDGS (Edwards et al. 2007). The higher price is based on (Hamelinck and Hoogwijk 2007). It is assumed that the price level of (Animal science Group 2005) is more accurate for the Dutch situation. d

2.8.4 Regional economic factors In Table 2.13, regional specific economic factors used in our calculations are summarized. The time horizon is assumed to be the same as the lifetime of the perennial crop and the economic lifetime of an ethanol production plant (i.e. 20 years). Although the cost of capital is generally higher for private industries, in this study the interest rate is set at 5.5 % (de Wolf and van der Klooster 2006). This figure is similar to the discount rates of between 5-6%, that have been used in other studies (Styles et al.; Bullard 2001; van den Broek et al. 2002; Rosenqvist and Dawson 2005). Since there is some debate regarding the appropriate value for discount rates, and results are very sensitive to this parameter, a sensitivity analysis is undertaken. Because of high agricultural productivity and the many competing uses for land in the Netherlands, land prices are relatively high (Eurostat 2008; AgriHolland 2009) (see Table 2.13). Costs for labour include taxes and insurance and are given according to the official wage rates set for Dutch agriculture as agreed between employers and unions (Collective Labour Agreement; CAO). In reality, most Dutch farmers do not integrate the cost of their own labour, family labour, or the cost of machinery in choices regarding crop planning and management (Poppe 2004; de Wolf and van der Klooster 2006). Nevertheless, we include these costs in order to compare intensive and less intensively managed crops. Although the land values are not generally included in crop budgeting, we incorporate them here to account for the opportunity cost of land use.

70

2.

Potential , spatial distribution and economic performance of regional biomass chains; the North of the Netherlands as example

Table 2.13: Regional cost parameters. Regional cost parameter a

Unit

Reference

Time horizon

y

20

Discount rate

%

(de Wolf and van der Klooster 2006; Houtsma 2008)

Land costs

€/ha

5.5 574715

(de Wolf and van der Klooster 2006)

Labour farmer

€/h

19.30

(de Wolf and van der Klooster 2006)

Labour farmer employee

€/h

12.08

(de Wolf and van der Klooster 2006)

Labour logistics

€/h

19.51

(NEA 2004)

Red diesel (agricultural use)

€/l

0.77

(Statline 2008d)

Diesel (transport use)

€/l

1.091b

(Statline 2008d)

Electricity

€/kWh

0.192

(Statline 2008a)

Heavy oil

€/l

0.36

(Statline 2008c)

Natural gas

€/m3

0.27

(Statline 2008a) b

Petrol €/l 1.471 (Statline 2008d) lifetime plantation and plant b Average petrol price in 2006 at an oil price level of 62 US$ per barrel. Includes excise duty 0.371 €/l (diesel) and 0.676 €/l (petrol) (Statline 2008b), 19% VAT, and 0.135 €/l margin (Shell 2008). a

71

72

3.

Spatial variation of environmental impacts of regional biomass chains

3

Spatial variation of environmental impacts of regional biomass chains F. van der Hilst, J.P. Lesschen, J.M.C. van Dam, M. Riksen, P.A. Verweij, J.P.M. Sanders, A.P.C. Faaij Renewable and Sustainable Energy Reviews (2012) Volume 16, Issue 4: 2053-2069

73

SHADES OF GREEN

ABSTRACT In this study, the spatial variation of potential environmental impacts of bioenergy crops is quantitatively assessed. The cultivation of sugar beet and Miscanthus for bioethanol production in the North of the Netherlands is used as a case study. The environmental impacts included are greenhouse gas (GHG) emissions (during lifecycle and related to direct land use change), soil quality, water quantity and quality, and biodiversity. Suitable methods are selected and adapted based on an extensive literature review. The spatial variation in environmental impacts related to the spatial heterogeneity of the physical context is assessed using Geographical Information System (GIS). The case study shows that there are large spatial variations in environmental impacts of the introduction of bioenergy crops. Land use change (LUC) to sugar beet generally causes more negative environmental impacts than LUC to Miscanthus. LUC to Miscanthus could have positive environmental impacts in some areas. The most negative environmental impacts of a shift towards sugar beet and Miscanthus occur in the western wet pasture areas. The spatially combined results of the environmental impacts illustrate that there are several tradeoffs between environmental impacts: there are no areas were no negative environmental impacts occur. The assessment demonstrates a framework to identify areas with potential negative environmental impacts of bioenergy crop production and areas where bioenergy crop production have little negative or even positive environmental impacts.

74

3.

Spatial variation of environmental impacts of regional biomass chains

3.1 Introduction In recent years, sustainability issues related to bioenergy have increasingly become a point of discussion in scientific, public and political arenas. At several levels, initiatives for sustainability criteria, codes of conduct and protocols have been developed to deal with this issue (Projectgroep 'Duurzame productie van Biomassa' 2006; Commission of the European communities 2008; Fehrenbach et al. 2008; Gallagher 2008; ISCC 2010). The steps towards sustainable biomass certification system are well described by Lewandowski and Faaij (2006) and van Dam et al. (2008; 2010b). However, sustainable bioenergy criteria are currently hardly implemented in (inter-) national policies. In addition, initiatives regarding formulation of codes of conduct are mainly developed in a general and top down approach and do not account for the significance of regional variation in adoption and implementation. In the scientific arena, sustainability of bioenergy has been an important research objective. In Blottnitz and Curran (2007) several studies regarding environmental impacts of bioethanol production chains have been reviewed. In most of the studies, energy balances and greenhouse gas (GHG) balances of production chains have been assessed. In others, such as Kaltschmitt et al. (1997), Lewandowski and Heinz (2003) Hamelink and van de Broek (2005), Malca and Freire (2006), Blottnitz and Curran (2007) and Styles and Jones (2007b) more complete Life Cycle Assessments (LCA) have been performed. Van den Broek et al. (2000a), Smeets et al. (2008; 2010) Rowe et al. (2009) and Börjesson (1999a; 1999b) are examples of studies on the sustainability of several bioenergy crops in specific countries that included a broad scope of environmental impacts and some even includes socio-economic impacts. The national scope of these latter assessments implies a relative aggregated level of analysis without differentiating for region specific conditions. Physical and socio-economic conditions often vary strongly between and within regions and determine the design of bioenergy chains: the chains need to be physically feasible in the geographical region, but also need to be compatible with the socio-economic context. In addition, the interactions between the bioenergy chain and the regional conditions have a strong influence on the actual environmental and socio-economic performance of bioenergy chains. A national approach could therefore give a distorted impression of the sustainability of bioenergy chains within regions. For that reason, the regional level is a more suitable level of analysis of the impact of biomass chains in order to safeguard sustainability of bioenergy chains for certification purposes. This is illustrated by van Dam (2009b) who made a region specific assessment of impacts of bioenergy production in the Pampa region in Argentina. Although this study took differences in physical context into account, the impacts were not assessed spatially explicit. Regional averages on physical 75

SHADES OF GREEN

parameters were used e.g. for soil quality or water availability. As the spatial variation in determining parameters of impacts of biomass chains is significant, impacts would preferably assessed spatially explicitly. The aim of this paper is to develop a methodological framework to assess the environmental impacts of regional biomass chains in a spatially explicit and integrated way. The production of ethanol from Miscanthus (Miscanthus x Giganteus) and sugar beet (Beta vulgaris L.) in the North of the Netherlands (the province of Groningen, Friesland and Drenthe) is used as case study. The case area was selected because of the high data availability. This case study was also used to analyse the economic viability of regional biomass chains (Van der Hilst et al. 2010). This study will explore the spatial variability of environmental impacts of bioenergy chains and builds on the previous analysis. The environmental impacts identified by several (inter-) national initiatives regarding sustainability criteria of bioenergy production were used as a starting point for a selection of relevant impacts and suitable indicators, as these criteria reflect the areas of concern for large scale bioenergy application (EC 2009a; NEN 2009; RSB 2009). The environmental impacts taken into account in this study are: GHG emissions (both during lifecycle and due to land use changes), soil quantity and quality, water use and water quality, and biodiversity. The selection and adaptation of methodologies to assess the environmental impacts is based on an extensive literature review on impact assessments. In this study, it is assumed that there is no competition with food production as the land for bioenergy crops is limited to the arable land that could become available by means of more efficient agricultural production. The development in land availability for bioenergy crops in the Netherlands is based on the study of de Wit et al. (2010).

3.2

Methods

The level of spatial detail of the analysis depends on the availability and resolution of the required input data. For several impacts, a postal code area was selected as the level of 2 spatial analysis. For other impacts, a grid level of 0.1-1km was selected. Because of the heterogeneity of the criteria, it is neither possible nor appropriate to draw generic system boundaries and select a single functional unit. Therefore, system boundaries and functional units were decided on for every individual indicator. In this study, it is assumed that only agricultural land (both arable land and pastures) is used for the cultivation of bioenergy crops. The effects of land use changes (LUC) for energy crop production were compared with the impacts of current agricultural land use. 76

3.

Spatial variation of environmental impacts of regional biomass chains

Nature areas were excluded beforehand. This is in line with the EU directive on the promotion of the use of energy from renewable sources (EC 2009a). The availability of agricultural land is therefore directly related to the intensification of agricultural land. As in the Netherlands the current state of agriculture is very advanced and consequently yield gaps are relatively small, only a small proportion of agricultural land is expected to become available for energy crops (de Wit and Faaij 2010; Van der Hilst et al. 2010). Since current fertilizer and pesticide application levels in the Netherlands already meet legislative limits, it is assumed that the slight production growth of the remaining agricultural land are not a result of increased inputs but from improved management only. It is therefore assumed that this intensification will not result in significantly more environmental impact. Consequently, the environmental impacts related to the productivity increase of the remaining agricultural land were not included in this assessment. The impacts of cultivation of Miscanthus and sugar beet were assessed for the entire agricultural area. This allows a straightforward interpretation of the analysis and provides a rationale for the selection of most appropriate areas energy crop production. Only potential direct effects of implementation of bioenergy chains were assessed. Second order effects and feedback loops, e.g. the effect of changes in nutrient leaching on biodiversity, are hard to quantify and to allocate and are due to their complexity, beyond the scope of this study. Indirect effects and displacement effects due to LUC were tackled beforehand by assuming that only surplus agricultural land could be dedicated to bioenergy crops. Based on the first order results of the assessment of the spatial variability of environmental impacts combined with results of the study on the spatial variation in economic viability of bioenergy chains (Van der Hilst et al. 2010), areas can be identified where bioenergy crop production is economically viable and have the least negative or possibly positive environmental impacts. The environmental impact analysis of LUC to bioenergy crops on a spatially explicit and detailed level requires large amount of spatial data of soil characteristics, climate, current land use, yield, crop management, fertilizer and manure inputs and deposition levels, see e.g. (Smeets et al. 2008; van Dam et al. 2009b). To deal with the large amounts of spatial data, a framework was developed to be able to calculate the impacts included in this study. The Miterra model used in our study, is derived from the MITERRA-Europe model, which was developed to assess the effects and interactions of policies and measures in agriculture on N losses on a regional level in EU-27 (Velthof et al. 2009). Miterra simulates 77

SHADES OF GREEN

the N and P balance, emissions of NH3, N2O, NOx, and CH4, leaching of N, NO3 concentration in groundwater, and changes in soil and biomass carbon stocks at a postal code level. For the application in this study, the model was adapted and input data and parameter values were made region specific. The main input data of the model are crop areas and livestock numbers, which were derived from the Geographical Information system Agrarian Businesses database (GIAB) on postal code level. For the assessment of the impacts not included in the Miterra model, appropriate methods were selected and adapted for spatial explicit application. In the following sections, the methods used to asses every individual impact are discussed.

3.2.1 GHG emissions One of the main drivers of bioenergy is the reduction of GHG emissions. Therefore, the GHG emission related to bioenergy production is an important indicator to take into account. The GHG emission of the bioenergy chains was calculated by combining the GHG emissions related to direct LUC and the GHG emissions during all stages of the ethanol production chain. GHG during life cycle GHG is emitted during the lifecycle of biofuels and is related to cultivation of the energy crop (diesel for agricultural machinery, seed pesticide and fertilizer production and fertilizer application), storage and transport (heat for drying, diesel for transport, dry matter loss) and processing (energy and chemical inputs for conversion process). The GHG emissions related to the production of bioenergy were calculated using a LCA approach which is often applied on bioenergy production chains. The ISO 14040-14049 guidelines articulate a preference for the use of system expansion in order to account for co products. As for legislative purposes energetic value based allocation is preferred in the EU directive on biofuels (EC 2008), this was applied in this study. The calculation of the GHG emissions was based on data provided by JEC (2008) on the production of biofuels. For cultivation parameters, universal data were substituted by region specific data provided in van der Hilst et al. (2010). In Appendix I (3.6), the data used to calculate the primary energy requirements and the GHG balances are depicted. GHG emissions due to LUC GHG emissions due to LUC are caused by changes in soil carbon stocks, above and below ground biomass and residues. When land is converted from one land use to another, carbon can accumulate (carbon sequestration) or diminish (carbon emissions). In addition, LUC causes changes in N2O emissions due to changes in fertilizer and manure application and drainage of organic soils. As it is assumed that there are no changes in livestock, livestock related N2O and methane emissions are assumed to remain constant when land 78

3.

Spatial variation of environmental impacts of regional biomass chains

is converted to energy crops. The Miterra model used here makes use of the methodology proposed in the IPCC guidelines to calculate the GHG emissions due to LUC (IPCC 2006). In Appendix I (3.6), the parameters used to calculate changes in GHG emissions due to LUC are presented. Direct N2O emissions from managed soils related to different N sources (manure, grazing, mineral fertilizer, crop residues and cultivation of organic soils) and indirect N2O emissions due to N leaching and N deposition were calculated by Miterra using emission factors of the IPCC guidelines (IPCC 2006; Velthof et al. 2007). The N surplus was calculated from the total N input (manure, mineral fertilizer, deposition and N fixation), the removal via crop harvesting and the N losses by gaseous emissions and runoff. Part of the N surplus is leached which results in direct N2O emissions (Velthof et al. 2009). In Appendix I (3.6.3), the parameters to calculate the changes in N2O emissions are presented. Due to lower nutrient requirements, less manure will be applied on herbaceous energy crops. As no change in manure production is assumed, a manure surplus could be developed. This potential manure surplus is not considered here as it is outside the scope of this study. The annual changes in SOC, the change in carbon flux from organic soils and the changes in annual N2O emissions were combined to calculate the changes in annual GHG emission.

3.2.2 Soil quality As soil quality is a broad and wide-ranging concept, it was narrowed down to indicators that are relevant for the region and for measuring the effect of bioenergy crop production on the soil. Soil organic matter (SOM) and soil erosion hazard are the indicators of soil quality that were selected in this study in line with the studies of Smeets et al. (2009b) and van Dam et al. (2009a; 2009b). Soil organic carbon As the SOC content is closely related to the amount of SOM, SOC is used as a proxy indicator. In addition to the benefits of carbon sequestration in the soil (see section 2.1.2), SOC is also linked to other important functions of the soil like water holding capacity, nutrient retention and soil structure (Kuikman et al. 2003; Rowe et al. 2009). SOC is therefore considered to be the most prominent indicator for soil quality (Reeves 1997). SOC is affected by soil type, climate, past and present land use, soil and water management (Romkes and Oenema 2004). In the Netherlands, most soils have a high SOC content. Pastures usually contain more SOC than arable land due to continuous land cover, higher root biomass, high animal manure application rates and absence of tillage. Sandy soils have generally lower organic matter content than clay soils because of higher 79

SHADES OF GREEN

organic decomposition rates (Romkes and Oenema 2004). As wet conditions slow down decomposition rate, soils with high water tables contain generally more SOC (Romkes and Oenema 2004). Peat lands store the largest amount of SOC but annual losses due to cultivation and drainage are high (Romkes and Oenema 2004). Changes in SOC were evaluated for the LUC related GHG. The methodology applied is discussed in section 3.2.1. Soil erosion In areas prone to soil erosion, LUC may have a significant effect on the actual erosion risk. The main on-site problem caused by erosion on agricultural land is the loss of fertile top soil which leads to degradation of arable soils (Pimentel et al. 1995; USDA and NRCS 1996; USDA and NRCS 2002; Romkes and Oenema 2004) and crop damage caused by abrasion or burial of seedlings or plants and the exposure and loss of seed (Riksen and de Graaff 2001). In addition, the transport of minerals, organic matter, residues and pesticides could cause contamination of surrounding surface water (USDA and NRCS 1996; van Kerckhoven et al. 2009). Furthermore, airborne particles due to wind erosion could affect human and animal health, machinery and infrastructure (USDA and NRCS 1996; 2002). The main factors determining the actual erosion risk, are the soil characteristics (especially soil moisture and soil structure), vegetation (soil cover) and slope (in case of water erosion) (Pimentel et al. 1995; Romkes and Oenema 2004). It is therefore important to assess the effect of LUC or farm management on these factors in regions with high erodible soils and periods with high erosive rainfall or wind. Most studies regarding erosion focus on erosion caused by water, as this is the dominant form of erosion in most areas (Pimentel et al. 1995). In the Netherlands, soil erosion due to runoff is mainly occurring in the sloping areas in the South-East (Eppink and Spaan 1989; Romkes and Oenema 2004). The erosion hazard map of the Netherlands by Eppink and Spaan (1989) shows no erosion risk due to rainfall runoff in the North of the Netherlands as this area has an outspoken flat morphology and the soils have a relatively high infiltration capacity. For this reason, rainfall related erosion is not included in this study. Wind erosion, however, does play a role in this region. Areas underlain by sandy and reclaimed cut-over peaty soils with frequent tillage on large fields without barriers are susceptible to wind erosion (Eppink and Spaan 1989; Riksen and de Graaff 2001; USDA and NRCS 2002; Romkes and Oenema 2004). It is estimated that the vulnerable area of wind erosion in the in the North Netherlands amounts up to 60 kha. Several methods have been developed to model wind erosion for different temporal and spatial scales, functionalities (specific circumstances) and impacts (soil loss, particle concentration). Ideally, soil erosion is continuously measured in the field or estimated by very exact wind erosion models fed with continues data on physical parameters. However, 80

3.

Spatial variation of environmental impacts of regional biomass chains

this is very time and capital intensive. In this study, the wind erosion equation (WEQ) method is applied. This is a relatively simple methods that requires less detailed input data and can be applied on a regional level as demonstrated by Van Kerckhoven (2009). The -1 -1 WEQ estimates the average soil loss due to wind erosion (in ton ha yr ). Equation 3.1 and Table 3.1, give the parameters of the WEQ derived from USDA and NRCS (2002), Morgen (2005) and van Kerckhoven (2009). The complex relations between the parameters and the input data to calculate the soil loss are explained in Appendix I (3.6.2).

E = ∫(IKVCL) Equation 3.1 Table 3.1: Parameters of the WEQ equation. Parameter E I

Erosion Soil erodibility index

K

Soil surface roughness factor

C

Climate factor

L

Length of field

V

Vegetation factor

Unit Potential average annual soil loss The potential annual soil loss per hectare for an open unsheltered isolated levelled field with a smooth bare crustless surface. Related to texture, organic matter, calcium carbonate concentration Reduction of the potential wind erosion due to ridges of the soil. Related to ridges and cloddiness made by field operations Local climatic erosivity. Related to wind speed, precipitation and temperature Unsheltered distance across a field along the prevailing wind direction Equivalent vegetative cover. Kind, amount and orientation of the vegetation expressed in Small Grain equivalents (SGe)

See Appendix 3.6.2

Ref

T ha-1·yr T ha-1·yr

Table 3.4

a,b

factor

Table 3.6

b

factor

Table 3.5

c

Table 3.6

d

m T ha-1·yr

a

(USDA and NRCS 2002) (van Kerckhoven et al. 2009) c (Heijboer and Nellestein 2002) d (NRCS 2006) b

When land is converted from conventional use to bioenergy crops, most factors included in the WEQ remain constant. The soil surface roughness factor could change slightly (±20%) due to modifications in tillage and planting practices. The main factor that will change is the equivalent vegetative cover. The erosion risk is determined for every individual crop for every month of the year. To calculate the erosion risk for rotations, the risk of the rotation crops are combined in a weighted summation based on the portions of the crops in the rotation. For grass, it is assumed that pastures are reseeded every 10 years. For Miscanthus a lifetime of 20 years is assumed (1 year of establishment, 19 years 81

SHADES OF GREEN

of full crop cover, early spring harvesting). All parameters used for the calculations of wind erosion are depicted in Appendix 1 (3.6.2). In line with the WEQ guidelines, the most critical month has been selected based on calculations (see Figure 3.12, Appendix I, section 3.6.3).

3.2.3 Water use and water quality Water use and water quality are often addressed in proposed sustainability criteria for bioenergy (Cramer 2007; Fehrenbach et al. 2008; NEN 2009). As a relatively small proportion of the total water consumption for biofuel production is used during the processing stage (<10%)(Berndes 2002; de Fraiture and Berndes 2009; Gerbens-Leenes et al. 2009; Dornburg et al. 2010), only the water use related to crop production is taken into account. The assessment of the impact on water quality is limited to evaluation of the effect of change in use of fertilizers on the water quality. Water use The change from current land use to energy crops may change the water balance of an area due to changes in evapotranspiration, runoff and percolation (Smeets et al. 2009b). The amount of water lost through evapotranspiration depends on crop type, growth stage, climate, soil characteristics growing period and agronomic practice (Brouwer and Heibloem 1986; Berndes 2002; Bessembinder et al. 2005; Dornburg et al. 2010). As in the North of the Netherlands only a minor proportion of the agricultural land is irrigated (1% in wet years to 11 % in dry years (Hoogeveen et al. 2003)), water withdrawal for irrigation is not incorporated in this study. Preferably, the impact of energy crops on fresh water availability for other functions is assessed on a water basin level (Dornburg et al. 2010). This approach however, requires detailed knowledge and data about the hydrologic flows within a specific water basin. In this study, the changes in the water balance when current 2 land use is converted to bioenergy crops was assessed on a spatial level of 1km grids. In addition, a comparison was made between the region specific water use efficiency (WUE) of the two bioenergy crops. To assess the potential water depletion due to the introduction of bioenergy crops, a simple water balance was made by comparing the evapotranspiration to the effective precipitation like done in the studies of Smeets and Faaij (2010) and van Dam et al. (2009b) (see Equation 3.3 in Appendix I 3.6.3). The calculation method of water depletion is made time and spatially explicit by making use of spatial data on long term averages of 2 monthly precipitation (1 km ) of the Royal Netherlands institute for Meteorology (KNMI). The effective precipitation (EP) is defined as the rainfall that is useful or usable in any phase of the crop production (Dastane 1978) and is derived from the actual rainfall making use of the USDA formula in the CROPWAT 8.0 model (FAO 2009a). The spatial 82

3.

Spatial variation of environmental impacts of regional biomass chains 2

distribution of monthly reference evapotranspiration (ET0) was constructed on a 0.1 km grid level for the North of the Netherlands by means of aggregation and spatial interpolation of daily ET0 data of the KNMI weather stations in the region. Although the Penman-Monteith equation is the most widely applied method to calculate ET0, the KNMI uses the Makkin method for practical reasons (Buishand and Velds 1980; Hooghart and Lablans 1988). The crop evapotranspiration coefficients (Kc factors) mainly depend on crop type, growth stage of crop and climate (Allan et al. 1998). The crop specific decadal Kc factors for the 2 Dutch situation were derived from the KNMI (2002). The spatial 0.1km grid maps of monthly ET0 were converted to land use specific monthly evapotranspiration maps, using the crop specific monthly Kc factors. The average monthly evapotranspiration of rotations were calculated using a weighted summation of the individual crops. As water deficits only occur during the summer in this area (KNMI 2002), the assessment of the change in total water deficit focussed on the months April-August. The parameters used for the calculation of the spatially explicit water deficit are depicted in Appendix I (3.6.3). The Water Use Efficiency (WUE) indicator is frequently applied in bioenergy related studies, like in Berndes et al. (2002), Dornburg et al. (2010), van Dam et al. (2009b), Fraiture and Berndes (2009) and Smeets and Faaij (2010). In this study, the WUE is used as a second indicator of water consumption. It provides an indication about the water requirements per unit crop produced, whereas the water deficit methodology only provides figures for water use per hectare. In order to assess the spatial explicit WUE of bioenergy crops, knowledge on water availability (precipitation and ground water) and the effect of different growth limitation factors on evapotranspiration rates should be considered. As here are too many uncertainties in these parameters, only a regional average WUE was calculated. Water quality The use of agricultural chemicals like pesticides and fertilizers may contaminate ground and surface water. A change in the use of these chemicals due to LUC in favour of energy crops is expected to have impact on the water quality. In this study, we focused on contamination of water due to fertilizer and manure use only, as these cause the most persistent problems in The Netherlands. An oversupply of N is likely to cause leaching of N to ground and surface water. The N surplus was calculated in the Miterra model as described in introduction section of the methods. Part of the surplus is denitrified and part will leach the ground and surface water. The amount of N leaching was calculated by multiplying the N surplus by a leaching fraction, which was based on the soil type, land use and groundwater level (Fraters et al. 2007). The NO3 concentration in ground water was 83

SHADES OF GREEN

calculated by dividing the N leaching flux by the precipitation surplus. For P there is no direct relation between the P accumulation in the soil and the concentration in ground and surface water. However, when P application levels exceed the P removal by crop uptake over longer periods of time, the risk of P leaching to water will increase. The P balance is calculated by Miterra making use of data on P inputs (manure, grazing, mineral fertilizer application and deposition) and P uptake by crops (Velthof et al. 2007). The input data required for the calculations of changes in NO3 concentration in groundwater and change in P balances are included in Appendix I (3.6.3).

3.2.4 Biodiversity In most of the developed sustainability criteria for bioenergy, impacts on biodiversity have been identified as an important area of concern. LUC is a strong driver of changes in biodiversity (Sala et al. 2000; UNEP 2002; Foley et al. 2005; Reidsma et al. 2006). Because of the loss, modification and fragmentation of habitats, the (indirect) expansion of agricultural land for energy crop production is perceived to be a major threat for biodiversity (Sala et al. 2009). Biodiversity is also affected by LUC related depletion, degradation and pollution of ecosystems and invasive species (Foley et al. 2005; Groom et al. 2008). Changes in habitats due to energy crop production are most significant when natural areas are converted to (intensive) agriculture areas (Schlegel et al. 2007). Most bioenergy sustainability criteria deal with this issue by proposing process indicators, e.g. by referring to national regulations and by excluding protected areas and land identified as area with high biodiversity from bioenergy production (EC 2008). In this study, these areas are excluded beforehand and only changes in agricultural land use are assessed. Currently, few guidelines are available about quantitative result indicators and methods to assess the impacts of energy crop production on biodiversity (Cramer 2007). This is especially true for assessing agro-biodiversity on a regional level. Over the last decades, several indicator systems have been developed to assess changes in biodiversity. These indicator systems vary to a great extent according to scale (global, national, regional or local), purpose (policy targets), and focal area of biodiversity (species, genetic variation, population size or ecosystems). The impact of energy crop cultivation on biodiversity depends on both local scale effects (choice of crop, management intensity, vegetation structure, substituted land use) and landscape scale effects (geographical location, scale and distribution of crops) (Eggers et al. 2009). The objective in this study is to quantify the impact of bioenergy crop production on agro-biodiversity on a local scale and identify the areas with a high risk on biodiversity loss. Two indicator systems, the High Nature Value (HNV) and the Mean Species Abundance (MSA) were selected for this assessment.

84

3.

Spatial variation of environmental impacts of regional biomass chains

HNV has become one of the main indicators to assess the impact on biodiversity of the 2007 – 2013 Rural Development Programs (EC 2009b) and for the integration of environmental concerns into the Common Agricultural Policy (Schlegel et al. 2007). High Nature Value (HNV) farmland are certain types of farmland that, because of their characteristics, can be expected to support high levels of biodiversity of conservation concern (EC 2009b). JRC and EEA developed a method to identify HNV farmland (Paracchini et al. 2008). Elbersen and van Eupen (2008) adapted the HNV method for the Netherlands based on more detailed spatial data on HNV parameters. The HNV indicators used in this study are explained in Appendix I (3.6.4). Based on the characteristics of the HNV types and characteristics of the energy crops, the impact of a shift to bioenergy crop production on biodiversity in a HNV area were qualitatively assessed on a detailed geographical level. The risk level of biodiversity loss was identified based on the extent to which the new land use function could meet the habitat functions requirements of the important species identified by the HNV cluster indicators (see Table 3.10). The areas where the implementation of Miscanthus or sugar beet could cause a ‘high risk of biodiversity loss’, ‘considerable risk of a biodiversity loss’ and ‘both positive and negative effects on biodiversity’ were mapped. The HNV approach does not provide information about the impact of a shift towards bioenergy crops on biodiversity outside the HNV areas. In addition, the HNV approach does not measure the change in biodiversity quantitatively. The Mean Species Abundance (MSA) is a quantitative indicator for change in biodiversity. It does not reflect individual species responses but represents the average response of the total set of original species relative to their abundance in undisturbed ecosystems (Alkemade et al. 2009). This indicator does not cover all aspects of the complex concept of biodiversity, but it can be used appropriately to assess changes in biodiversity due to changes in land use for bioenergy crops. It was successfully applied in several global en regional studies concerning changes in biodiversity (MNP 2006; Secretariat of the Convention on Biological Diversity and Netherlands Environmental Assessment Agency 2007; Dornburg et al. 2010). The MSA values for several land cover types are (included in Table 3.12, of Appendix I) was linked to a map of current land use. However, as this land use map does not account for diversity within a land use type, management practices or locally available biodiversity, the MSA provides little differentiation for agricultural land in the Netherlands. For all areas outside the HNV areas, the MSA indicator is applied to assess the impact of LUC to bioenergy crops on biodiversity.

85

SHADES OF GREEN

3.2.5 Overview methods and integration of results Figure 3.1 shows an overview of the bioenergy supply chain, the impacts that are assessed, the methods that are applied and the type of results that are produced in this study.

Figure 3.1: Overview of the bioethanol chain and the environmental impacts assessed, the indicators used and the models applied. The bottom boxes indicate the type of result that is generated from the assessments.

For an integral picture, the GIS maps layers of the potential environmental impacts have been combined. As all impacts are measured in different units and scales, combining impacts requires standardisation of the GIS map layers. For all indicators, maximum standardisation was applied in which all negative impacts were translated to a scale from 0 to -10 and all positive effects to a value between 0 and +10. In this standardisation process, impacts equal to the impact of current land use were set to ‘0’, all increases in impacts compared to current land use were assumed to be negative and all decreases in impacts were assumed to be positive. In order to prevent a skewed distribution in standardised values due to extreme results of the indicators, the maxima were set on 2 86

3.

Spatial variation of environmental impacts of regional biomass chains

times the standard deviation from the average. In order to compare the results of Miscanthus and sugar beet, the same standardisation has been applied for the results of both crops. In this integrated assessment, it is assumed that all environmental impacts are equally important.

3.3 Results 3.3.1 GHG emissions GHG during lifecycle The GHG emissions over the production chain of bioethanol are substantial (see Figure 3.2). The results found in this assessment are comparable with the figures found in previous studies (Elsayed et al. 2003; Bergsma et al. 2007; Hamelinck and Hoogwijk 2007; Hoefnagels et al. 2010). Bioethanol production of Miscanthus causes fewer GHG emissions nd than first generation biofuels. As thermal energy requirements for 2 generation conversion are assumed to be fuelled by part of the feedstock, the process requires less fossil energy. Figure 3.2 shows the relative importance of GHG emissions due to changes in SOC compared to the emission during the lifecycle. The uncertainty bars indicate the spatial variation in SOC change for a 20 year period. As the emission reduction potential of a bioenergy production chain is highly related to the location of production this needs to be spatially assessed (see Figure 3.3).

Figure 3.2: GHG emissions due to SOC changes and during lifecycle of bioethanol production from Miscanthus and Sugar beet. For comparison, the average GHG emissions of petrol over the lifecycle are depicted.

GHG emissions due to LUC Figure 3.4 shows the changes in GHG emissions due to LUC from current land use to Miscanthus and sugar beet. This includes the changes in carbon stocks (Figure 3.3), the changes in carbon fluxes from organic soils and N2O emissions (Figure 3.11 in Appendix II 87

SHADES OF GREEN

3.7.1). When grassland is converted to Miscanthus, a carbon loss due to land clearing (biomass carbon, up to 5.4 ton C/ha) and soil preparation prior to Miscanthus planting (soil carbon, 1 t C/ha on sandy soils, 2.5 t C/ha on clay soils) is compensated by relatively large above and below ground biomass stock (up to 17.5 t C/ha) that Miscanthus achieves compared to grass. When cropland is converted to Miscanthus, relatively large amounts of carbon are sequestered both in biomass and soil carbon (8.5 t C/ha on sandy soils, 21.5 t/ha on clay soils). For Miscanthus, the GHG emission reduction through carbon -1 sequestration varies between 0 to -78 kg GJ ethanol. For sugar beet, the GHG emissions due -1 to changes in carbon stocks vary between 0 and 54 kg GJ ethanol. These figures are in line with the studies of Foereid et al. (2004) and Schneckenberger and Kuzakov (2007) on sequestration potential of Miscanthus and with the study of Kuikman (2003) on SOC in the Netherlands.

Figure 3.3: Δ SOC when current land use is converted to Miscanthus (left) or to sugar beet (right).

LUC related N2O emissions contribute significantly to the GHG emissions. Conversion of current land use to Miscanthus generally decreases N2O emissions. Especially when pastures (generally high fertilizer and manure application levels) are converted to Miscanthus, N2O emissions decrease significantly (±1.7 ton CO2-eq per ha/y). For Miscanthus, the GHG emission reduction though LUC related N2O emission reduction varies between 0 and -21 kg CO2-eq GJ-1ethanol. LUC to sugar beet cause an increase of N2O emissions in almost all areas, with highest increases on arable land. This is mainly caused by the higher fertilizer application levels and the increased emissions from crop residues (sugar beet leaves and crowns). For sugar beet ethanol, the LUC related N2O -1 emissions contribute 0 -15 kg GJ ethanol to the GHG balance. See Figure 3.11 in Appendix II 3.7.1. Because LUC could cause an increase in one type of GHG but a decrease in another GHG, changes could balance each other out (for example: leaving the residues of sugar beet on 88

3.

Spatial variation of environmental impacts of regional biomass chains

the field will increase SOC but it will also increase N2O emissions from residues). In general, LUC towards Miscanthus will reduce GHG emissions, especially conversion of arable land. LUC to sugar beet will generally cause a net increase in GHG emissions. The amount of GHG emissions is relatively high in areas now in use for pastures. For Miscanthus, the GHG emission reduction compared to current land use varies between 0 -1 -1 to -159 kg CO2-eq GJ ethanol (-15 ton CO2-eq ha ). When arable land is converted to sugar beet, relatively small changes in GHG emissions occur. The GHG emissions due to LUC to -1 -1 sugar beet range from 0 to 148 kg GJ ethanol (-1 ton CO2-eq ha ).

Figure 3.4: Δ GHG when current land use is converted to Miscanthus (left) or to sugar beet (right).

3.3.2 Soil erosion -1

The risk on erosion (in ton ha·y ) for current land use is shown Figure 3.13 in Appendix II, section 3.7.2, and the changes in the risk on erosion due to LUC are depicted in Figure 3.5. The Figures are based on the parameters for decreased soil moisture levels and relatively high wind velocities (see Appendix I, 3.6.2). -1

Although the overall risk on erosion is relatively small (max < 6 ton ha·y ), this could still have severe negative effects. In the current situation, light sandy soils in the south east of the region (veenkoloniën) are the areas most at risk for soil erosion. In the areas prone to soil erosion, relative large changes in erosion risk occur. When arable land is converted to Miscanthus, erosion risk is significantly reduced. Although it is assumed that grass is renewed every 10 years and Miscanthus every 20 years, risk on erosion increases when pastures are converted to Miscanthus. This is related to the planting date of grass which is generally seeded in September. By the time the soil gets prone to erosion in spring, grass already covers the soil to a large extent. Miscanthus, however, needs to be planted in spring and has a slow crop development in the initial growth stage. Therefore, in the year of establishment the risk on erosion is relatively high. These findings are in line with Kort 89

SHADES OF GREEN -1

-1

et al. (1998). The risk on erosion increases up to 3 ton ha y when sugar beet replaces crop rotations. When sugar beet replaces pastures, the risk on erosion can even increase -1 -1 up to 9 ton ha y on sandy soils.

Figure 3.5: Δ Wind erosion risk in kg soil/ha/y when current land use is converted to Miscanthus (left) or to sugar beet (right).

3.3.3 Water use and water quality Water consumption In the current situation, evapotranspiration of crop rotation and pasture land exceed the effective precipitation in the months April to September (see Figure 3.14, Appendix II). Therefore, a cumulative deficiency occurs during these months. As the effective precipitation exceeds the evapotranspiration considerably during the rest of the year, all water shortages are eventually replenished. However, temporary shortages during the summer could cause damages to agricultural production or natural areas. The evapotranspiration of Miscanthus exceeds the evapotranspiration of both pastures and rotations of arable crops from May to September. During July and August the evapotranspiration of sugar beet exceeds the average evapotranspiration of rotations and pastures. Therefore, LUC to bioenergy crops could cause higher water consumption rates during summer. As water tables are regulated artificially in the Netherlands, more water should be supplied to the areas with higher water requirements. The developments in crop specific evapotranspiration over time in relation to the effective precipitation are depicted in Figure 3.14 in Appendix II. The spatial variation in water deficits during the months April-August caused by evapotranspiration of current land use is depicted in Figure 3.15 of the Appendix. The largest water shortages occur in the western part, as precipitation is relatively low in this

90

3.

Spatial variation of environmental impacts of regional biomass chains

area compared to the rest of the region. The change in water shortages when current land use is converted to Miscanthus or to sugar beet, are depicted in Figure 3.6. In addition to the spatial water balance, the regional WUE was measured. In Figure 3.16 of the Appendix, the WUE for Miscanthus and sugar beet is depicted. Most studies indicate that the WUE of C4 crops is generally higher than the WUE of C3 crops (Berndes 2002; Dornburg et al. 2010). However in this study, it was found that the WUE of Miscanthus (2.23 g/kg water) is lower that for sugar beet (3.87 g/kg water). This can be explained by the fact that the competitive advantage of the C4 photosynthesis pathway decreases significantly further away from the equator. In addition, as climatic conditions and agricultural management are optimal for sugar beet production, it achieves relatively high WUE. The WUE found for Miscanthus is in line with the 2.1 g/kg found by Clifton Brown (2000) and the range of 1-9.5 g/kg provided by Berndes (2002). The regional WUE of sugar beet is in line with the range of 2.55-3.83 g/kg given by (Dornburg et al. 2010).

Figure 3.6: Δ Water depletion during summer (mm) when current land use is converted to Miscanthus (left) or sugar beet (right).

Water quality The Miterra model was used to calculate the nitrogen (N) and phosphorus (P) balance in the soil. Based on the N balance and the water surplus, the NO3 concentration in the ground water was calculated. The maps of the N and P balance are included in Appendix II, F17 and F18. As N and P fertilizer application levels are closely linked, the spatial patterns of changes in N and P balances are quite similar. Because of the large differences between the current fertilizer and manure input on pasture and the fertilizer requirements of Miscanthus, the N and P surplus will be significantly reduced when pasture area is converted to Miscanthus. Also in the areas where sugar beet substitutes grassland N and P surpluses are reduced. However, when sugar beet substitutes a rotation of crops, N and P 91

SHADES OF GREEN

surpluses increase. The high increases of N and P occur when rotations with high proportions of cereals are converted to sugar beet. This is mainly caused by the relatively high fertilizer application rates of sugar beet compared to other rotation crops. The differences in patterns between the N balance and NO3 concentrations can be explained by the differences in leaching rates (lower for grassland, peat soils and lower -1 ground water levels). The NO3 concentration in mg NO3 l is shown in Figure 3.7. The current NO3 levels are 72 mg/l on average and 75% of the zip code districts exceed the standards of 50 g/l of the Nitrate directive (Council of the European Communities 1991). In general, the NO3 concentration decreases when current land use is converted to -1 Miscanthus. The average change in NO3 concentrations is -53 mg l and varies between 0 -1 in areas with a large fraction of arable land and -187 mg l in zip code areas with a high proportion of grassland. When pasture area is converted to sugar beet, a decrease in NO3 -1 concentration (of max -75 mg l ) occurs. When arable land is converted to sugar beet, an -1 average increase in NO3 concentrations of 16.5 mg l is simulated.

Figure 3.7: Δ NO3 concentration (mg/l) when current land is converted to Miscanthus (left) or to sugar beet (right).

3.3.4 Biodiversity In Figure 3.8, the HNV areas were biodiversity is at risk when current land use is converted to bioenergy crops are mapped. It shows that in almost the entire pasture area in the north of the Netherlands which comprises high densities of important species (endangered bird species and large shares of European populations of meadow and/or wintering birds), biodiversity is at risk when current land use is converted to either Miscanthus or sugar beet.

92

3.

Spatial variation of environmental impacts of regional biomass chains 3

The HNV areas where biodiversity is at risk when converted to sugar beet (532·10 ha, of 3 which 193·10 ha at ‘high risk’) are more extensive than for conversion to Miscanthus (473 3 3 ·10 ha, of which 91·10 ha at ‘high risk’). The areas where LUC to bioenergy crops is expected to cause a considerable risk of biodiversity loss are considered to be areas where energy crops should be introduced with care. The areas with a high risk on biodiversity 3 loss are considered to be no-go areas for bioenergy crops. In some areas (59·10 ha) the introduction of Miscanthus could have both a positive and a negative effect on biodiversity. These are mainly the heterogeneous areas and the extensive managed arable areas in the east of the region.

Figure 3.8: Risk of biodiversity loss when current land use is converted to Miscanthus (left) or to sugar beet (right) based on HNV cluster indicator.

For the areas not included in the HNV areas, the change in biodiversity is expressed in the change in MSA value. This mapped in Figure 3.19 of Appendix II. The change in MSA value indicates an increase in biodiversity when pasture land is converted to Miscanthus and especially when arable land is converted to Miscanthus. The MSA values indicate a decrease in biodiversity when pasture land is converted to sugar beet cultivation and a status quo in biodiversity when sugar beet is cultivated in arable areas. The results show that sugar beet will induce a risk of biodiversity loss both within and outside the HNV areas and that the only areas where no negative effects on biodiversity occur are the areas already in use for intensively managed arable crops.

3.3.5 Integrated results In Figure 3.9 the integrated results of all environmental impacts are spatially depicted for both Miscanthus and Sugar beet. The results are expressed as an average score on a standardised scale from -10 to +10. In general, Miscanthus has better overall scores (mean = 2.33) than sugar beet (mean = -0.85). The results discussed in previous sections show a 93

SHADES OF GREEN

wide distribution of scores for the individual impacts. As the integrated results show average scores close to 0, it illustrates that there are several trade-offs between impacts and that positive and negative scores balance each other out to some extent.

Figure 3.9: Integrated result of environmental impact of a shift from current land use to Miscanthus (left) or sugar beet (right) based on standardised maps of individual environmental impacts.

There are no areas where only positive impacts occur when current land use is changed to sugar beet or Miscanthus. For both crops, the western wet pasture area and the peaty south east area of Drenthe appear to be the areas with most negative impacts. The conversion of pasture land to arable crops on (reclaimed) peat areas has negative impact on carbon stocks, risk on erosion and biodiversity. These negative impacts are relatively small for Miscanthus but quite severe for sugar beet. The impacts in these areas would even be more severe in case the pasture areas were not as intensively managed as they are today. Sugar beet has also a negative environmental performance in the eastern part of the region. This area is characterised by arable land on sandy soils. As sugar beet is a more intensively managed crop than average arable crop rotations, negative impacts on the NO3 concentration, P balance, soil erosion and also biodiversity are relatively large. For the same reasons, Miscanthus (extensively managed crop) scores relatively well in this area. Sugar beet has the least negative impacts in the northern pasture areas on clay soils and negative impacts are partly balanced out by reduced risk of nutrient leakage and water consumption. Comparing the integrated results of the environmental impacts assessed in this study with the results of the economic performance of sugar beet and Miscanthus in this region assessed in the study of Van der Hilst et al. (2010), show that the areas with best environmental performance are also the areas with the least cost of production (in €/GJbiomass). These are the areas currently in use for intensive agriculture. However, in 94

3.

Spatial variation of environmental impacts of regional biomass chains

most of these areas bioenergy crops cannot compete with current land use as conventional crops have a better economic performance on these suitable soils. Based on the results of the environmental impacts and economic performance, an area of 24 kha can be identified as most promising for Miscanthus as in these areas it could compete with current land use, it has relatively low production costs and has the least negative and also positive environmental impacts. This well exceeds the proportion of agricultural land that could become available for bioenergy crop production in 2015 (9.6 kha arable land and 1.5 kha pasture land in the considered region) as indicated in the Refuel study (de Wit and Faaij 2010). As the study of Hilst et al. (2010)showed that sugar beet cannot compete with current land use and has relative high production costs in the whole agricultural area of the North of the Netherlands and this study shows mainly negative environmental impacts, it is assumed that sugar beet not to be a promising bioenergy crop in this region.

3.4 Discussion and conclusions Preferably, impacts would be assessed on most detailed spatial level. However, several input parameters of the Miterra model are only available on a postal code level. In addition, crop areas on a zip code level provides a proper indication of the crop mix in rotation, whereas a land cover recording of a single year on a detailed grid level ignores the effect of rotating crops. Furthermore, as there are uncertainties in several key input parameters such as crop management, leakage fraction per soil type, water surplus etc, impact assessment on a detailed grid level could give a too optimistic impression on the accuracy of the results. There are some drawbacks of the Miterra model: The fertilizerderived N2O emissions are calculated in accordance with the Tier 1 approach of the IPCC. However, as indicated by Smeets et al. (2009a) and Lesschen et al. (2010) this neglects the variability of emissions due to differences in environmental conditions, management and crop type. The Miterra Model includes only regional averages of climate data and in the calculation of leaching of nutrients and decomposition of organic matter, water surplus of precipitation was included but water table levels were neglected. Despite these shortcomings, the Miterra model was able to combine several spatial datasets in a consistent way and therefore provides excellent information on the relative differences in impacts on a spatially detailed level. As indicated by Hoefnagels et al. (2010), Cherubini et al. (2009) and van Dam et al. (2010a), GHG balances of biofuels are very sensitive for assumptions regarding system boundaries, allocation methods and input parameters. Therefore, the results of the GHG assessment should be interpreted with care.

95

SHADES OF GREEN

High soil losses due to wind erosion occur only incidentally when all relevant conditions are disadvantageous at the same time. By using an average monthly climate factor in the WEQ, no accurate indication is provided on how often erosion thresholds are met. Although the WEQ is not the most sophisticated approach to assess the risk on erosion, it does provide a method to discriminate between crop covers and to assess the spatial variability in the erosion risk on a regional level. It is recommended to use actual field measurements and more exact models to spatially assess erosion risk on a regional level in order to achieve more accurate figures on the exact amount of soil loss. The cumulative water deficit is assessed taking the spatial-temporal precipitation and evapotranspiration levels into account. It neglects however groundwater level, flow schemes of the water basin and the extensive water table regulation system. Therefore, the spatial water balance provides too little information to assess where actual droughts or water surpluses will occur. In order to assess the actual effect on water tables it is recommended that more advanced hydrologic models are applied. In this study the HNV and MSA indicators were applied to identify areas with a high risk on biodiversity loss. The HNV approach enables to exclude the most vulnerable areas from a biodiversity perspective, which in line with the bioenergy sustainability criteria such as developed by RSB, RSPO, NTA8080. However, due to a lack of information on the response of species on the introduction of bioenergy crops, no quantitative assessment could be made. The MSA indicator on the other hand, enables to quantify the overall effect of a shift to bioenergy crops on biodiversity but because the MSA values does not account for diversity within a land use type, management practices or locally available biodiversity, the MSA provides little differentiation for agricultural land in the Netherlands. In order to assess the impact on biodiversity in a quantitative way, we conclude in line with Rowe (2009) that more empirical data are required on the impact of specific bioenergy crops on the occurrence of individual species. The calculations of the environmental impacts are all based on a very broad range of input parameters. As the results are affected by the assumptions made on climate, soil, agronomy and management applied, they should be interpreted with care. However, as the assumptions are consistently applied, it provides a good indication the relative spatial differences in environmental impacts. As all impacts are related to ecosystem functionality, they are heavily interlinked with each other. In some cases, impacts reinforce each other and in other cases, tradeoffs between impacts occur. Although it is hard to prove causality and to quantify relations, it would be interesting to assess the spatial correlation between several impacts.

96

3.

Spatial variation of environmental impacts of regional biomass chains

All environmental impacts of Miscanthus and sugar beet cultivation have been integrated in a standardised map of overall environmental performance. This required an interpretation of the dose-response relationship: to what extend is the caused effect a negative or positive impact. Both the dose-response relationship and the thresholds for actual damage are expected to be spatially heterogeneous, time related, and ,in some cases, contradictive. In addition, cause-effect relationships could be linear, quadratic, exponential, inversely proportional, etc. As there is too little knowledge available of the dose-response relationship, a simple approach of maximum linear standardisation with no threshold was applied in this study. More expertise is required on the translation of a change in value of an indicator to an actual impact. This requires not only knowledge on physical relationships but also on the focal points of policy, the willingness to accept and the different points of view of the stakeholders involved. In the integration of environmental impacts all impacts were considered to be equally important, whereas ideally multiple perspectives from stakeholder on the hierarchy of environmental impacts would be included in a spatially explicit Multi Criteria Analysis (MCA). This first explorative step could be further developed to an integrated assessment method by means of a stakeholder assessment and more advanced standardisation and weighting methods. The integrated assessment shows that there are no areas in the North of the Netherlands where only positive effects occur when bioenergy crops are introduced and that there are several trade-offs between the impacts. In general, sugar beet causes relatively many negative environmental impacts especially in pasture areas. In these areas, the LUC -1 related GHG emissions are up to 148 kg GJethanol and the risk on soil erosion increases to 9 -1 ton ha . Also, there is a high risk on biodiversity loss in these areas. On the other hand, a -1 shift towards sugar beet could cause a decrease of 75 mg l in the NO3 concentration of groundwater and a decrease in the seasonal water deficits with 100 mm. LUC to Miscanthus has less negative environmental impacts and in the areas currently in use as arable land many positive environmental impacts are expected to occur. In these areas, -1 the LUC related GHG emissions are reduced by -159 kg GJethanol , the risk on soil erosion -1 -1 could be reduced with 4 ton ha , the NO3 concentration is reduced with 53 mg l , and the shift towards Miscanthus could have a positive contribution to biodiversity. On the other hand, at some locations, the seasonal water depletion increases with 150mm when arable land is converted to Miscanthus. For both crops, the western wet pasture areas appear to be the area with most negative impacts. It can be concluded that the spatial correlation between impacts and current land use is higher than between impacts and soil type. The combination of the spatial distribution of the environmental impacts and the economic performance of sugar beet and Miscanthus assessed in the study of Van der Hilst et al. (2010) shows that the areas with best environmental performance are also the areas with the least cost of production (in €/GJbiomass).Considering the large amount of negative 97

SHADES OF GREEN

impacts on areas currently in use as pastures, it could be recommended to exclude these areas for bioenergy crop production. In addition, as sugar beet causes relative many negative impacts, large scale introduction of this crop for bioenergy purposes should be carefully considered and implemented wisely. This assessment has resulted in understanding in which areas the introduction of bioenergy is likely to cause severe environmental impacts and in which areas little negative and even positive effects are caused. This study well illustrates the high spatial variability of environmental impacts of potential bioenergy chains and therefore it stresses the importance to assess the sustainability of bioenergy supply chains in an integrated and spatially explicit way. This study provides a set of methodologies to assess the spatial variation of environmental impacts of bioenergy chains. Therefore, it contributes to identify promising locations for bioenergy production from an environmental point of view. Because of the generic characteristics, the methodologies and tools applied for this study can also be used for other case studies. It should be noted that this type of spatial explicit regional assessment on environmental impact of bioenergy chains cannot replace site specific environmental impact assessment.

3.5 Acknowledgements This study is part of the Climate changes spatial planning program and is funded by the Dutch government, the European commission and Shell. The authors gratefully acknowledge the contribution of Saskia Visser (wur, soil centre) and Geert Sterk (landscape ecology, geo-informatics and hydrology; Utrecht University) for their expertise and information on (wind) erosion; Rob Sluiter, Raymond Sluiter and Jan Willem Nooteboom of the KNMI for their help to collect the required climatological data; Wilbert van Rooij, Marc van Oorschot, Rob Alkemade of PBL for sharing their knowledge about Mean Species Abundance and assessment methods; Berien Elbersen and Michiel van Eupen for their input on High Nature Value areas in the Netherlands; Iris Lewandowski for her knowledge and expertise about Miscanthus; Jan den Besten and Henk Norel for their expertise on water and water related problems in Hunze and Aa in Groningen; and Roland van Zoest of the GeoDesk Wageningen University for his help on GIS mapping of the climatological parameters.

98

3.

Spatial variation of environmental impacts of regional biomass chains

3.6 Appendix I: Methods and Input data In this Appendix additional information is provided on the method applied and the input data used in this study. Every section is dedicated to an environmental impact assessed in the study. In Appendix 2 (3.7), additional information is provided on the results of this study.

3.6.1

Parameters for calculation of the GHG emissions

In this study, the GHG emissions during the lifecycle of ethanol production and GHG emissions due to land use change (LUC) were considered. The GHGs taken into account in this study are CO2, N2O and CH4 and are expressed in CO2-equivalents. A 100 year time horizon is considered for the global warming potential of the GHGs, based on the fourth assessment report of the IPCC (2007c). The general LCA approach is applied making use of region specific data on cultivation (Van der Hilst et al. 2010) and more general data on conversion (JEC, 2008). Additional data used for the calculation of the GHG during the life cycle are depicted in Table 3.2. The Miterra model was used to calculate the GHG emissions due LUC. The Miterra model is based on the method proposed by the IPCC (2004). The IPCC default values were applied in this study, making use of region specific data where appropriate and available. Some studies indicate that it could take up to 50-100 years to reach a new SOC equilibrium (Kuikman et al. 2005). However, in this study a time horizon of 20 years is assumed in line with the IPCC (2006) and as proposed by the EC (2008) and NTA 8080 (NEN 2009). For peaty soils, GHG emissions were calculated using an annual flux instead of changes in SOC stocks. In Table 3.2 the factors used to calculate the SOC changes, carbon fluxes and N2O emission are presented.

99

SHADES OF GREEN

Table 3.2: Energy requirements and emissions related to inputs and field activities during crop cultivation.

Seed a

Fertilizer b

Pesticide c

Energy e

Chemicals

Input Miscanthus rhizomes grass seeds sugar beet seed summer wheat winter wheat barley Sugar beet rape seed maize Seed potato N fertilizer P2O5 fertilizer K2O fertilizer herbicide insecticide fungicide diesel petrol fuel oil coal natural gas electricity steam CaO H2SO4

Unit MJ/ha MJ/ha MJ/ha MJ/ha MJ/ha MJ/ha MJ/ha MJ/ha MJ/ha MJ/ha MJ/kg MJ/kg MJ/kg MJ/kg MJ/kg MJ/kg MJ/MJ MJ/MJ MJ/MJ MJ/MJ MJ/MJ MJ/MJ MJ/MJ MJ/kg

Energy requirement 1600 290 551 389 425 291 138 15 110 1890 45.33 15.25 8.88 265 214 173 1.11 1.16 1.09 1.05 1.02 2.96 1.13 5.03

Unit kg CO2-eq/ha kg CO2-eq/ha kg CO2-eq/ha kg CO2-eq/ha kg CO2-eq/ha kg CO2-eq/ha kg CO2-eq/ha kg CO2-eq/ha kg CO2-eq/ha kg CO2-eq/ha kg CO2-eq/kg kg CO2-eq/kg kg CO2-eq/kg kg CO2-eq/kg kg CO2-eq/kg kg CO2-eq/kg kg CO2-eq/MJ kg CO2-eq/MJ kg CO2-eq/MJ kg CO2-eq/MJ kg CO2-eq/MJ kg CO2-eq/MJ kg CO2-eq/MJ kg CO2-eq/MJ

GHG emission 110 20 48 40 43 30 12 2 10 104 5.33 0.71 0.45 5.37 5.37 5.37 0.08 0.09 0.09 0.11 0.06 0.13 0.07 0.79

MJ/kg

3.92

kg CO2-eq/MJ

0.19

NH3

2.32 MJ/kg 39.95 kg CO2-eq/MJ N-hexane MJ/kg 1.25 kg CO2-eq/MJ 0.10 a Energy requirements and related emissions for seed/rhizome/cuts production are derived from (Biewinga and Bijl 1996; Kaltschmitt and Reinhardt 1997; Bullard and Matcalfe 2001; West and Marland 2002) b Energy requirements and related emissions for fertilizer production (Kaltschmitt and Reinhardt 1997; Bullard and Matcalfe 2001; West and Marland 2002). The N2O emissions related to N fertilizer application on the field are based on the Tier 1 method of the IPCC (IPCC 2006). c Energy requirements and emissions related to pesticide production vary to a great extend as indicated by (Helsel 1997). However, averages for herbicides, insecticides and fungicides as indicated by (Kaltschmitt and Reinhardt 1997; Bullard and Matcalfe 2001; West and Marland 2002) are used in this study. d Based on (Kaltschmitt and Reinhardt 1997). e Derived from (JRC 2009).

100

3.

Spatial variation of environmental impacts of regional biomass chains

Table 3.3: Values used to calculate the change in carbon stocks when land is converted for growing energy crops. Values are based on the IPCC (2006) unless otherwise indicated. Variable Climate region a

Specification warm temperate, moist

Symbol

Value

Error

Dimension

Soil types

sandy soils

SOCref

71

NA

t C ha-1

clay soils

SOCref

95

NA

t C ha-1

cropland

EF

10

0.9

t C ha-1yr-1

grassland

EF

2.5

0.9

t C ha-1yr-1

long term cultivated permanent grassland

Flu

0.69

0.12

factor

Flu

1

NA

factor factor

organic soils Land uses b

cropland grassland

Management Input c

Yield d

Harvest index

e

Carbon contentf

cropland

full tillage

Fmg

1

NA

grassland

improved grassland

Fmg

1.14

0.11

factor

cropland

high input + manure

Fi

1.44

0.13

factor

grassland

high input

Fi

1.11

0.07

factor

grassland

medium input

Fi

1.00

NA

factor

Miscanthus

Regional average

Yld

13.7

+0.23/-0.35

t dm ha-1 yr-1

sugar beet

Regional average

Yld

65

+0.09/-0.43

t fresh ha-1 yr-1

grass

Regional average

Yld

11.5

+0.34/-0.29

t dm ha-1 yr-1

Miscanthus

HI

0.40

+0.36/-0.19

sugar beet

HI

0.67

grass

HI

0.50

Herbaceous3

Cc/dm

0.47

index index

+0.3/-0.6

index t C t dm-1

crop land Cc/dm 0.47 t C t dm-1 Time needed to reach equilibrium soil Equilibrium C stock D 20 yr yr time g a According to the IPCC, the Netherlands is situated in the cold temperate moist climate region. However, the soil organic carbon concentrations of mineral soils and especially the carbon fluxes of organic soils measured in the field are more associated to the default values for a warm temperate moist climate. In addition, due to climate change the climate in the Netherlands is expected to become warmer in the next coming decades. Most parameters are equal for the warm and cold variations of the temperate climate. Only the soil carbon content of several soil types and the carbon fluxes of arable land and pastures on organic soils are different. b Perennial herbaceous crops like Miscanthus are not included as a separate land use category in the IPCC guidelines. Because it suits the characteristics of perennial grassland, it is assumed that the factors of improved grassland are most appropriate for Miscanthus cultivation, as is done for switchgrass by van Dam et al. (2009b). It is assumed that the management factor of improved grassland is the most suitable for miscanthus cultivation. When Miscanthus is cultivated on land previously in use as arable land, carbon stocks are expected to change as grass has to be removed prior to the planting of Miscanthus rhizomes, and as management and crop characteristics are different. In the year of the land use change, grass is removed and land needs to ploughed. Since in this year of establishment the management applied is similar to cropland, this year the carbon stock values for cropland are used. For the remaining part of the lifetime (19 years), it is assumed the carbon stock values for grassland are the most appropriate for Miscanthus cultivation. In the year of establishment, soil carbon stock decreased due to enhanced oxidation due to ploughing. The biomass carbon stock is re-established relatively quickly due to the high growth rate of Miscanthus. 101

SHADES OF GREEN c

Because of the relatively low fertilizer requirements of Miscanthus, the medium input factor of grassland is assumed to be the most appropriate factor for Miscanthus cultivation. d Yield levels are specific for the region, yield levels of sugar beet is derived from (PPO et al. 2006) of grass is derived from (Animal science Group 2005) and Miscanthus is based on (Christian and Riche 1999; Clifton-Brown et al. 2004; Elbersen et al. 2005). The presented yield levels are averages of the region. e Harvest index is based on the harvestable yield compared to the total above and below ground biomass. The root-to-shoot-ratio of Miscanthus fluctuates during life time but below ground biomass is expected to achieve a dry mass of 1.66 times the max above ground biomass (Himken et al. 1997; Neukirchen et al. 1999; Kahle et al. 2001). The Harvest index of sugar beet is based on (Kuikman et al. 2003). And also in line with assumptions of 65 ton sugar beet fresh weight, 23% dm, 5 ton leafs/ha and 10% beet residue. It is assumed beat leaves are left on the field and ploughed back into the soil. Harvest index of grass is based on (Kuikman et al. 2003). f Consistent with figures found in field trials of Miscanthus in Germany (Kahle et al. 2001), (Lewandowski et al. 2000) reported values of 0.48-0.5. g Consistent with guidelines of the European directive (Commission of the European communities 2008).

3.6.2 Parameters for calculation of the impact on soil quality Soil organic carbon and the risk on erosion have been selected as indicators for soil quality. Soil organic carbon was also calculated in order to assess the GHG emissions due to land use change. The parameters for this calculation are therefore included in section 3.6.1. The risk on erosion is related to multiple factors. In Equation 3.2 the wind erosion equation (WEQ) is depicted. It shows the complex relationship of all input variables. The steps to follow when applying the WEQ are depicted in Equation 3.2a to Equation 3.2k.

E = ∫(IKVCL) Equation 3.2

E1 = I Equation 3.2a

E= E1 ⋅ K 2 Equation 3.2b

E= E2 ⋅ C 3 Equation 3.2c

= C 34.48 ⋅

V

3

( EP )

2

Equation 3.2d 10

P 9 12  EP = 115∑ i =1    T − 10 

Equation 3.2e

E4 = ( F 0.3484 + E30.3484 − E20.3484 )

2.87

Equation 3.2f 102

3.

Spatial variation of environmental impacts of regional biomass chains

−0.3829  L  L   ⋅ exp  −3.33   F = E2 ⋅ 1 − 0.1218    L0    L0    

Equation 3.2g

L0 = 1.56 ⋅ 10 ⋅ E2 6

−1.26

⋅ exp ( −0.00156 ⋅ E2 ) Equation 3.2h

E 5= g ⋅ E 4 h

g= exp ( −0.759 ⋅ V − 4.74 ⋅ 10 ⋅ V + 2.95 ⋅ 10 ⋅ V −2

−4

2

3

Equation 3.2i

)

Equation 3.2j −2

−3

−5

h =1.0 + 0.893 ⋅ 10 ⋅ V + 8.51 ⋅ 10 ⋅ V − 1.5 ⋅ 10 ⋅ V 2

3

Equation 3.2k E5 I K C v P EP T L L0 V

Potential average annual soil loss the soil erodibility index soil surface roughness factor climate factor Annual average wind velocity precipitation Effective precipitation Temperature unsheltered distance across a field Maximum field length equivalent vegetative cover

T ha-1·yr T ha-1·yr factor factor ms-1 inches inches F m m SGe ton/ha

The potential average annual soil loss (E) calculated with WEQ is based on an annual climate factor (C) and field conditions (I, K, L, and V) during the critical wind erosion period of the year (USDA and NRCS 2002).The soil erodibility index (I) is the potential annual soil loss per hectare for an open unsheltered isolated levelled field with a smooth bare crustless surface (USDA and NRCS 2002). It is assumed that the soil erodibility factors of the Flemish soils found by van Kerckhoven et al. (2009), are also representative for similar soil types in the Netherlands. For soil types not included in the study of van Kerckhoven, like sandy-clay soils, the wind erodibility groups and indexes of the USDA and NRCS (2002) were applied (see Table 3.4). The soil roughness factor (K) indicates the reduction of the potential wind erosion due to ridges of the soil and varies for different crops. The K factors for crops in Flanders developed by Verbist, Schiettecatte and Gabriels, quoted by Van Kerckhoven et al. (2009), are expected to be suitable for the Dutch situation as Flemish and Dutch agricultural practices are similar (see Table 3.6). The climate factor (C) is related to precipitation, temperature and average wind speed (USDA and NRCS 2002). As 103

SHADES OF GREEN

spatial variation in precipitation and wind velocity in the North of the Netherlands is relatively small and the focus of this study is on relative differences between vegetation cover, one average climate factor for the whole region was applied in this assessment (see Table 3.5). The length of field factor (L) is related to the unsheltered distance along the prevailing wind direction (USDA and NRCS 2002). It would require extensive remote sensing or field work to assess the field length and the presence of potential wind barriers for every individual plot in the region. As the study of van Kerckhoven et al. (2009) shows that the final result is not sensitive for changes in field length, in this study it is assumed that the field length (L) equals the maximum field length (L0). The vegetation factor (V) indicates the reduction in potential wind erosion due to a vegetation cover. The reduction of wind erosion potential depends on the kind, amount and orientation of the vegetative material and is expressed in small grain equivalents (SGe) (USDA and NRCS 2002). The vegetation factor (V) changes over time during the growth season of crops. The SGe values for crops during a year are derived from the WEQ worksheet (NRCS 2006) and adapted to the Dutch situation using the planting dates for crops in the Netherlands (see Table 3.7). The orientation of the rows of crops (perpendicular or parallel to the prevailing wind direction) highly affects the erosion risk. In this study, it is assumed that the crop rows are perpendicular to the prevailing wind direction to minimise the wind erosion risk. Table 3.4: Erodibility indexes for soil classes in the Netherlands. Soil type

Erodibility index (I)a

Sand

384

Loamy sand

304

Loamy very fine sand

300

Light loamy sand

110

Sandy loam

74

Loam

82

Noncalceaous loam <20%clay

126

Clay

128

Heavy clay

188

ton/ha/y

a

Dune sand 695 Based on the erodibility indexes provided by van Kerckhoven et al. (2009) for the Flamisch soils and the soil erodibility groups of the USDA (2002).

104

3.

Spatial variation of environmental impacts of regional biomass chains a

Table 3.5: Climate parameters for calculation of monthly and annual climate factor.

Jan

Precipitation

Temperature

PE

Wind

Climate

inch/month

F

PE

m/s

factor

2.72

35.60

9.51

5.5

0.44

Feb

1.77

35.78

5.85

5.0

0.87

Mrt

2.41

40.82

6.79

5.2

0.73

Apr

1.74

45.50

4.02

4.6

1.44

Mei

2.26

53.42

4.32

4.1

0.88

Jun

2.86

57.92

5.01

4.0

0.61

Jul

2.87

61.70

4.64

3.9

0.66

Aug

2.23

61.70

3.50

3.7

0.99

Sep

2.83

56.30

5.15

3.8

0.49

Okt

2.75

49.28

5.99

4.1

0.46

Nov

3.07

41.90

8.55

4.9

0.38

Dec

2.95

37.76

9.54

5.3

0.39

Annually 21.99 48.14 72.86 4.5 0.59 a Based on long term averages 1971-2000 of the Royal Netherlands Institute of meteorology (KNMI 2002). The data is form the weather station in Eelde. This is the closest weather station to the wind erosion susceptible area. PE is effective precipitation.

105

SHADES OF GREEN

Kb

grass subsequente

Miscanthus first yearf

Miscanthus subsequent

fallow

0.00 0.00

grass firstd

0.00

maize

0.00 0.87

sugar beet

summer barley

0.87

industrial potatoes

winter wheat

0.00

seed potatoes

summer wheat

Jan

consumption potations

Month

Va

winter barley

Factor1

Table 3.6: Monthly vegetation factors, soil roughness factors and crop tolerance for blowing soil.

0.00

0.00

6.03

7.84

0.00

7.84

0.00

Feb

0.00

0.92

0.05 0.92

0.00

0.00 0.00

0.00

0.00

7.84

7.84

0.00

7.84

0.00

Mrt

0.32

1.28

0.79 1.28

0.15

0.00 0.00

0.00

0.00

7.84

7.84

0.00

7.84

0.00

Apr

2.79

2.29

3.85 2.29

0.18

0.03 0.04

0.12

0.05

7.84

7.84

0.00

7.84

0.00

Mei

5.39

4.82

5.78 4.82

0.53

0.38 0.27

2.58

0.94

7.84

7.84

0.39

7.84

0.00

Jun

5.98

5.98

5.98 5.98

7.88

4.55 3.33

12.65 5.10

7.84

7.84

3.79

7.84

0.00

Jul

5.98

5.93

5.93 5.25

14.61 6.10 13.60 14.87 8.31

7.84

7.84

7.55

7.84

0.00

Aug

5.11

4.48

4.48 4.48

14.87 3.01 14.87 14.87 8.58

7.84

7.84

7.84

7.84

0.00

Sep

4.48

3.48

4.48 3.48

14.87 0.78 14.87 14.87 3.83

0.07

7.84

7.84

7.84

0.00

Okt

4.48

0.37

4.48 0.37

8.96

0.78 4.42

8.64

1.46

2.16

7.84

7.84

7.84

0.00

Nov

4.48

0.73

4.48 0.73

0.78

0.78 0.78

2.80

1.46

6.03

7.84

7.84

7.84

0.00

Dec

4.48

0.87

4.48 0.87

0.78

0.78 0.78

2.80

1.46

6.03

7.84

7.84

7.84

0.00

year

3.62

2.67

3.73 2.61

5.30

1.43 4.41

6.18

2.60

6.27

7.84

4.24

7.84

0.00

year

0.77

0.77

0.77 0.77

0.66

0.66 0.66

0.81

0.77

0.81

0.81

0.81

0.81

0.81

ns ns ns 2.97 2.97 2.97 <1.23 4.94 ns ns ns ns Tolc T ha-1 ns a V factor is the vegetation factor. These are derived from (NRCS 2006) and adapted to the Dutch situation using the planting dates derived from (Schreuder et al. 2008). It is assumed that after harvest the stubble of cereals and the leaves of potatoes and wheat will not be ploughed under until the end of the year. this will provide a soil cover and therefore protection against wind erosion. Degradation of the residues is not taken into account. b K factor is the soil roughness factor. These values are derived from the study of van Kerckhoven et al.(van Kerckhoven et al. 2009)for the Flemisch situation. c Crop tolerance defined as the maximum wind erosion that a growing crop can tolerate without an economic loss to crop yield and/or quality (USDA and NRCS 2002). ‘ns’ means not susceptible. The values of Miscanthus are based on the values for the C4 crop Sorghum. d Year of establishment of grass. Grass is generally seeded in autumn providing cover in the winter. e Subsequent years of grass. In this study an average lifetime of 10 years is assumed for grassland. The maximum value of cover is assumed for grass (according to WEQ 7000lbs/ac SGe)(NRCS 2006). f The first year is the year of establishment. Miscanthus is planted in May. Growth is generally slow in the first months after establishment but with an assumed density of 20.000 plants per hectare it will provide sufficient cover. In March after the first growing season, Miscanthus will be mowed, but all stems and leaves are assumed to be left in the fields as there is too little economic incentive to recover them from the field. From the second year on harvest will take place in March and 30% of the yield (tops and leaves) will be left in the field which provided sufficient cover to protect against wind erosion.

106

3.

Spatial variation of environmental impacts of regional biomass chains

3.6.3 Parameters to calculate the impact on water use and water quality In this study, two indicators to assess the impact of bioenergy cropping on water quantity were used. The simple water balance used to calculate the water deficit based on local effective precipitation and local evapotranspiration and is depicted in Equation 3.3. The input variables used to calculate the water balance are presented in Table 3.7. The water use efficiency (WUE) indicator was used to express the water requirements per unit biomass and unit bioethanol. Regional averages of evapotranspiration and crop yield were used (see Table 3.8). The Miterra Model was used to calculate the impact on water quality (N and P surplus and NO3 concentration in ground water). The input data required to feed the model are enclosed in Table 3.9.

WSi = ∑ [ −((ET0 i ⋅ Kci ) − EPi )] Equation 3.3

EPi = ( Pi ⋅ (125 − 0.2Pi ) / 125) WS ET0 Kc EP P i

Total water shortage in month i Reference evapotranspiration of month i Crop evapotranspiration coefficient for specific growth stage in month i Effective precipitation in month i Precipitation in month i Month January to December

mm/month mm/month Factor mm/month mm/month

Table 3.7: Crop evapotranspiration coefficient of crops based on (KNMI 2002). Kc Initial stage

Mid season

Late season

Start

days

Growth stages Initial Crop phase develop

Summer wheat

0.30

1.15

0.25

12-feb

174

18

37

80

49

Winter wheat

0.40

1.15

0.25

24-sep

311

149

70

70

23

Crop

Growth season

Mid season

Late season

Potato

0.50

1.15

0.50

16-apr

175

33

39

55

33

Sugar beet

0.35

1.20

1.00

9-apr

215

42

60

74

37

Maize

0.30

1.20

0.50

23-apr

170

29

48

58

38

Miscanthusa 0.85 1.20 0.85 1-jan 365 83 91 140 51 a In line with Danalatos (2007), the Kc factor of Miscanthus is expected to be equal to the Kc factor of maize. However, as the threshold for photosynthesis of Miscanthus is 3-5 °C below the threshold of maize (Long et al. 2001), it is expected that the Kc factor of Miscanthus to increase faster in the beginning of the growth season and to be higher during the last stage of the growth season compared to maize. Because Miscanthus is dormant and turns brown during winter, it is assumed that evapotranspiration is zero during this period.

107

SHADES OF GREEN

Table 3.8: Regional average crop evapotranspiration (based on figures of Table 3.7) and yield of Miscanthus and sugar beet. Crop

Average evapotranspiration mm ha-1 y-1 613-671 386-426

Miscanthus Sugar beet

Average yield Odt ha-1 13.66 14.95

Miscanthus

Grass

Corn cob mix

Maize

Rape seed

Sugar beet

Industrial potato

Seed potatoes

Feeding potatoes

Winter barley

Summer barley

Unit

Winter wheat

Parameter

Summer wheat

Table 3.9: Values used to calculate the change in nitrogen related emissions and phosphorous balance when land is converted for growing energy crops.

a

kg/ha

71

98

55

73

103

68

101

148

95

104

87

200

13

P fertilizer b

kg/ha

6

8

6

7

6

4

7

13

8

3

11

15

2

N manure c

kg/ha

144

144

144

144

144

144

144

144

144

221

143

155

24

17

17

17

3.5

3.5

3.5

1.8

35

15

13.9

30

5

N fertilizer

N crop

d

kg/ton 17

Harvest ratio e kg/kg 0.67 0.67 0.67 0.67 0.69 0.69 0.69 0.69 0.25 0.86 0.86 0.5 0.37 a Based on the total fertilizer use derived from [ref!] and distributed over crop types groups according to fertilizer requirements. b Based on the crop requirements according to [ref!] minus the P application by manure c Based on total N requirement minus the N application by fertilizers d Derived from (Velthof and Kuikman 2000) e Derived from (Kuikman et al. 2003)

3.6.4 Parameters to calculate the impact on biodiversity Two different indicator systems were selected to assess the impact of land use change to Miscanthus and sugar beet on biodiversity: the High Nature Value (HNV) and the Mean Species Abundance (MSA) approach. The HNV approach based on the work of Elbersen and Van Eupen (2008) was applied. In Europe, and especially in the Netherlands, arable farming has often been intensified to the point where it can no longer be characterised as HNV. However, according to the EC (2009b) it is possible for more intensive farmland to continue to support important populations of species of conservation concern. Agricultural areas in the Netherlands host endangered bird species and large shares of European populations of meadow and/or wintering birds (Elbersen and van Eupen 2008). Elbersen and Van Eupen indentified three types of HNV farmland and for each type two cluster indicators were developed (see Table 3.10).

108

3.

Spatial variation of environmental impacts of regional biomass chains

Table 3.10: HNV types and related cluster indicators based on (Elbersen and van Eupen 2008) HNV type 1 Includes semi-natural vegetation within and outside protected natural if managed by extensive farmland practices 2

3

Includes farmland areas with a high density of wet (e.g. ditches) and/or green (e.g. tree lines) landscape elements, extensive farming practices and high number of farmland birds Includes farmland areas which are not included in type 1 or 2 but still host a large share of meadow and wintering bird populations

Cluster indicator A semi-natural vegetation outside protected natural areas B semi-natural vegetation inside protected natural areas C Extensively managed farmlands on dry sandy soils D Extensively managed wet peat land pasture areas

E F

Medium intensively managed peat land pasture areas Medium intensively managed arable land

Elbersen and van Eupen (2008) mapped all clustered indicators for the Netherlands on a 2 1km grid level. To what extent an area complies with the cluster indicator is expressed in the deviation of the national average score for the specific indicator. Only the areas that score at least one standard deviation above national averages on HNV cluster indicators were included in our assessment (score > average + 1SD). For the prevalence of meadow bird species however, all areas that score above national averages were included (score > average). The motivation to focus particularly on these species is provided by the fact that the Netherlands provides important habitats for the preservation of red list bird species and populations of European and world conservation concern. In Table 3.11, the expected impact of Miscanthus and sugar beet cultivation on the risk on biodiversity loss in HNV farmland area is described for every cluster indicator based on the findings of several studies. The expected impact of bioenergy crop production on biodiversity within a defined HNV area was linked to the spatial map of the HNV cluster indicators. This enabled the mapping of areas.

109

SHADES OF GREEN

Table 3.11: The expected impact of bioenergy crops on several HNV cluster indicators. Characteristics of HNV are based on Elbersen and van Eupen (2008). Risk of biodiversity loss caused by a shift towards bioenergy crops is based on several biodiversity studies. Cat

Characteristics HNV

Rationale for effect on biodiversity

Assumed effect of introduction bioenergy crop on risk of biodiversity loss Miscanthus Sugar beet

A + B

This category includes natural grassland, heather and dune grassland, salt marshes, grazed natural areas and these areas host species indicative for heather and peat. (Most of these areas are already filtered out by excluding natural areas.) This category includes extensive managed dry sandy grassland and arable land with high density of green landscape elements and host vegetation indicative for extensive managed agricultural land and birds requiring heterogeneous landscapes.

Environmental pressure, further habitat loss and biodiversity decline are expected when forest, grassland, peatland and wetlands are converted into (monoculture) plantations (Eggers et al. 2009). Changes in habitats due to energy crop production are most significant when natural areas are converted into (intensive) agriculture areas (Schlegel et al. 2007).

It is assumed that a shift to Miscanthus or sugar beet would cause a high risk on biodiversity loss in HNV areas A+B.

Spatial variation in physical or environmental conditions enhances the opportunity to fulfil sufficient life supporting functions for species (Shmida and Wilson 1985; Schouten et al. 2009). Due to the height of the crop (+/- 3m), Miscanthus may have a positive effect on landscape diversity (Eggers et al. 2009) and it could provide protection against predators and weather conditions during winter (EEA 2007; Bellamy et al. 2009). The height of the crop is likely to be disadvantageous for the population density of meadow birds (Bellamy et al. 2009). Wet landscape elements could remain disregarding the crop introduced. Drainage of wetland for conversion into biomass crops will have a negative effect on biodiversity (Schlegel et al. 2007). Conversion of grassland into cropland results in local extinction of most plant species and animals whose habitat is largely dependent on plant species composition (Sala et al. 2000; Schlegel et al. 2007). There is a strong relation between land use intensity and biodiversity decline (Lewandowski and Schmidt 2006; Reidsma et al. 2006; EEA 2007; Schlegel et al. 2007; Groom et al. 2008).

It is assumed that Miscanthus could have both a positive and a negative effect on biodiversity in HNV areas C+D

C + D

This category includes wet pasture areas, wet landscape elements, extensive pasture management, vegetation indicative for extensive managed agricultural land and birds requiring wet and heterogeneous areas.

110

It is assumed that sugar beet will cause a high risk on biodiversity loss in HNV areas C+D

3.

Spatial variation of environmental impacts of regional biomass chains

E

This category includes wet pasture areas that host rare species (including red list species) or host a high share of European or world population of species.

See Rationale nr. 5 See Rationale nr 8 and 9 It is assumed that in areas that host the highest amount of rare species and highest shares of important populations, no changes to land use could be made without diminishing biodiversity. Because of the importance of these populations from a conservation perspective, not only the areas with the highest densities but also the areas with above average densities of important species are included in this assessment.

It is assumed that Miscanthus and suagr beet causes a high risk of biodiversity loss in areas with a high density of important species.(HNV areas E) Miscanthus and Sugar beet are assumed to cause a considerable risk of biodiversity loss in areas with above average density of rare species

F

This category includes low or medium intensive arable areas that host rare species (including red list species) or host a high share of European or world population of species.

See Rationale 9 and 12

It is assumed that Miscanthus could have both a positive and a negative effect on biodiversity depending in HNV areas F

High risk biodiversity loss

Considerable risk biodiversity loss

It is assumed that sugar beet causes a considerabl e risk of biodiversity loss in HNV areas F

Both positive and negative effects

In addition to the HNV indicator, the MSA indicator was applied in the areas outside the HNV farmland areas. The MSA indicator is based on several drivers for changes in biodiversity: land cover change, land use intensity, fragmentation, climate change, atmospheric nitrogen deposition and infrastructure development (Alkemade et al. 2009). As this study focuses on short term LUC, only the effect of land cover change combined with land use intensity was taken into account. The other drivers are assumed to remain constant. All perennial energy crops are assigned a single MSA value and the same holds true for all arable crops.

111

SHADES OF GREEN

Table 3.12: Relevant MSA values derived from Van Rooij (2008) and Dornburg et al. (2008). Land cover type

MSA value

Primary forests

1.0

Natural grass & shrub lands

1.0

Light used primary forests

0.7

Secondary forests

0.5

Agro forestry

0.5

Extensive agriculture

0.3

Set aside land

0.3

Abandoned land

0.3

Perennial bioenergy crops

0.3

Forest plantations

0.2

Oil palm

0.2

Man made pastures

0.2

Irrigated intensive agriculture

0.05

Intensive agriculture

0.1

Oil crops and cereals

0.1

Built up areas

0.05

112

3.

Spatial variation of environmental impacts of regional biomass chains

3.7 Appendix II: Additional Results In this Appendix, results additional to the results shown in the main paper are described.

3.7.1 Additional results of GHG emissions In addition to the GHG emissions, the energy balance over the lifecycle is an important indicator. The results of the analysis of the energy requirements are presented in Figure 3.10. For comparison first generation bioethanol production from maize end wheat are also included. The energy use during the production chain of ethanol production is substantial. Especially the conversion of first generation bioethanol requires considerable energy inputs for steam production. Bioethanol production of Miscanthus requires less nd primary energy than first generation biofuels. As thermal energy requirements for 2 generation conversion are assumed to be fuelled by part of the feedstock, the process requires less fossil energy. Compared to bioethanol from wheat and maize, ethanol from sugar beet requires less primary energy and causes fewer GHG emissions during the lifecycle.

st

nd

Figure 3.10: Energy requirements for 1 and 2 the Netherlands.

generation bioethanol production in the North of

In Figure 3.4 of section 3.3.1, the GHG emissions due to LUC are depicted. As N2O is the main contributor to the total GHG emissions, the spatial variation in N2O emissions due to LUC are depicted here. In Figure 3.11 the change in N2O emissions due to a shift from current land use to Miscanthus and sugar beet is depicted. N2O emissions are generally high in areas with organic soils and in areas where high amounts of fertilizer/manure are applied. It should be noted that background emission related to manure management contribute to a large extent to N2O emissions. However, it is assumed that these remain constant when land uses are changed. Conversion of current land use to Miscanthus 113

SHADES OF GREEN

generally decreases N2O emissions. Especially when pastures (with generally high fertilizer and manure application levels) are converted to Miscanthus, N2O emissions decreases significantly with about 1700 kg CO2-eq per ha/y. An increased cultivation of sugar beet causes increases of N2O emissions in almost all areas. Highest increases in N2O emissions due to conversion to sugar beet occur on arable land. This is mainly caused by the higher fertilizer application levels and the increased emissions from crop residues (sugar beet leaves and crowns).

Figure 3.11: Δ N2O emissions when current land use is converted to Miscanthus (left) or to sugar beet (right).

3.7.2 Additional results of impact on soil quality In the analysis of the effect on the risk on erosion, the parameters for the most critical month need to be applied in the wind erosion equation (WEQ). Therefore, it was first required first to identify the most critical month for wind erosion in the region. In Figure 3.12 the erosion risk for every individual crop is presented for sandy and clay soils. The figures show that April is the most critical month in the Netherlands. This is due to decreasing soil moisture, relatively high wind velocities and insignificant soil cover of most crops. The risk on erosion for current land use is depicted in Figure 3.13. Especially arable land on sandy soils in the east of the region are the areas most at risk.

114

3.

Spatial variation of environmental impacts of regional biomass chains

Figure 3.12: Erosion risk for individual crops on sandy soils (left) and clay soils (right) over the year.

Figure 3.13: Risk on erosion in kg soil/ha/y current land use.

3.7.3 Additional results of impact on water quality and quantity In order to determine to what extent a shift to bioenergy crops affects the water deficit, the water balance of current land use needed to be determined. Therefore, the evapotranspiration of all individual crops over the entire year was calculated. The development of evapotranspiration over time in relation to the effective precipitation specific for one location (weather station Eelde) is depicted in Figure 3.14. In spring and during the end of the growing season the evapotranspiration of grass exceeds the evapotranspiration of crop rotations. However, midsummer, during the peak of the growth season, the evapotranspiration of arable crops is higher than the evapotranspiration of pastures. There is little difference between the evapotranspiration of rotations on sand and on clay soils; the mixture op crops balances out the differences in individual crop evapotranspiration. The evapotranspiration of Miscanthus exceeds the evapotranspiration of both pastures and rotations of arable crops from May to September. The evapotranspiration of sugar beets is lower than rotation’s average in the 115

SHADES OF GREEN

beginning of the growth season, however during July and August the evapotranspiration of sugar beet exceeds the average evapotranspiration of rotations. As the effective precipitation exceeds the evapotranspiration considerably during the rest of the year, all water shortages are replenished over the year. However, temporary shortages during the summer could cause damages to agricultural production or natural areas. Therefore, the cumulative water deficits during the drier months (April-Sept) were assessed in this study. As water tables are regulated artificially in the Netherlands, elevated crop evapotranspiration causes higher requirements of water deliverables. In Figure 3.15, the water depletion in the months April-September for current land use is depicted. The largest water deficits occur in the western part of the region. This is due to the lower precipitation in this area. However, as this is close to the IJsselmeer, water deficits can easily be replenished. In addition to the water balance, the WUE was used as an indicator. In Figure 3.16, the water use per gram biomass and per gram ethanol is depicted. Although C4 crops are generally more water efficient, sugar beet requires less water per unit biomass and unit ethanol than Miscanthus in this area.

Figure 3.14: Effective precipitation and crop specific evapotranspiration development over the year for the climate characteristics of Eelde (weather station in Groningen).

The Miterra model was used to calculate the N and P balance in the soil. The maps of the nitrogen and phosphorous balance are depicted in Figure 3.17 and Figure 3.18. As N and P fertilizer application levels are closely linked, the spatial patterns of changes in N and P balances are similar. Because of the large differences between the current fertilizer and manure input on pasture and the fertilizer requirements of Miscanthus, the N and P surplus will be significantly reduced when pasture area is converted to Miscanthus. Also in the areas where sugar beet substitutes grassland N and P surpluses are reduced. However, when sugar beet substitutes a rotation of crops, N and P surpluses increase. The high increases of N and P occur when areas with high proportions of cereals are converted to 116

3.

Spatial variation of environmental impacts of regional biomass chains

sugar beet. This is mainly caused by the relatively high fertilizer application rates of sugar beet compared to other rotation crops.

Figure 3.15: Water depletion (mm) during summer (April-September) for current land use.

Figure 3.16: WUE for Miscanthus and Sugar beet biomass (left) and bioethanol (right).

3.7.4 Additional results for impact on biodiversity In Figure 3.19 the HNV areas that are at a considerable or high risk of biodiversity loss are excluded. For the remaining area, the change in MSA value for a conversion from current land use to bioenergy crops is mapped.

117

SHADES OF GREEN

Figure 3.17: Δ N balance in soil when current land use is converted to Miscanthus (left) or to sugar beet (right).

Figure 3.18: Δ P balance in soil when current land use is converted to Miscanthus (left) or to sugar beet (right).

Figure 3.19: Change in MSA value for LUC to Miscanthus (left) and to sugar beet (right) outside HNV areas. 118

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

4

Spatiotemporal land use modelling to assess land availability for energy crops – illustrated for Mozambique F. van der Hilst, J.A. Verstegen, D. Karssenberg, A.P.C. Faaij Accepted for publication in Global Change Biology Bioenergy (in Press). Doi: 10.1111/j.17571707.2011.01147.x

121

SHADES OF GREEN

ASTRACT A method and tool have been developed to assess future developments in land availability for bioenergy crops in a spatially explicit way, while taking into account both the developments in other land use functions, such as land for food, livestock and material production, and the uncertainties in the key determinant factors of land use change. This spatio-temporal land use change model is demonstrated with a case study on the developments in the land availability for bioenergy crops in Mozambique in the timeframe 2005-2030. The developments in the main drivers for agricultural land use, demand for food, animal products and materials were assessed, based on the projected developments in population, diet, GDP and self-sufficiency ratio. Two scenarios were developed: a business-as-usual (BAU) scenario and a progressive scenario. Land allocation was based on land use class-specific sets of suitability factors. The land use change 2 dynamics were mapped on a 1 km grid level for each individual year up to 2030. In the BAU scenario, 7.7 Mha and in the progressive scenario 16.4 Mha could become available for bioenergy crop production in 2030. Based on the Monte Carlo analysis, a 95% confidence interval of the amount of land available and the spatially explicit probability of available land was found. The bottom-up approach, the number of dynamic land uses, the diverse portfolio of land use change drivers and suitability factors, and the possibility to model uncertainty mean that this model is a step forward in modelling land availability for bioenergy potentials. 122

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

4.1 Introduction Dedicated bioenergy crops are assumed to be the main contributors to future bioenergy supplies (Smeets et al. 2007; Dornburg et al. 2010). Therefore, land availability for energy crop production is one of the key determining factors for bioenergy potentials. As competition for land and related indirect land use changes (iLUC) are to be avoided (IPCC 2011), the land available for bioenergy crops depends on the land required for other land use functions. In recent years, an increasing number of studies have been published on bioenergy potentials on a global (e.g. Berndes et al. 2003; Hoogwijk et al. 2005; Smeets et al. 2007; Dornburg et al. 2010), European (e.g. Ericsson and Nilsson 2006; EEA 2007; Fischer et al. 2007; de Wit and Faaij 2010), national (e.g. Faaij et al. 1998; van den Broek et al. 2001; Walsh et al. 2003; Sang and Zhu 2011) and regional level (e.g. van Dam et al. 2009a). However, most of these studies have assessed biomass potentials on a spatially aggregated level. The disadvantage of such studies is that they provide only limited information on the location of the land available for bioenergy crops. Potential yield levels and environmental and socio-economic impacts of energy crop production are strongly related to the physical and socio-economic conditions of a location (van Dam et al. 2009a; 2009b; Van der Hilst et al. 2010; Beringer et al. 2011; Van der Hilst et al. 2012a); therefore, it is important to assess where land is (or could become) available for bioenergy production. Land use changes (LUC) result from complex interactions between human and biophysical driving forces that act over a wide range of temporal and spatial scales (Verburg et al. 1999). Several methodologies and models have been developed to simulate and explore land use change LUC (Veldkamp and Lambin 2001). These models differ in terms of scale (e.g. regional, global), process (e.g. deforestation, urbanisation), discipline (e.g. economic, environmental), approach (e.g. extrapolating historical trends, driving forces) and complexity (e.g. methods, resolution). A review of several land use models is provided by Agarwal et al. (2001) and Verburg et al. (2004). The Integrated Model to Assess the Global Environment (IMAGE) is an example of a framework that models land use change on a global level (Alcamo et al. 1998; MNP 2006). However, the global modelling level, the aggregated modelling approach, and the low number of both dynamic land use types and allocation factors makes it less suitable for regional or national assessments. Lapola et al. (2010) used the LandSHIFT model to simulate land use change on a national level in order to assess indirect land use changes and related carbon emissions for a fixed biofuel production target in Brazil for 2020. However, due to the low resolution and the limited number of both dynamic land use classes and allocation factors, this type of modelling is 123

SHADES OF GREEN

less suitable for spatially detailed analyses of multiple dynamic land use types. The Conversion of Land Use and its Effects (CLUE) modelling framework was developed in 1996 and has progressively been improved since then (CLUE-s and Dyna-CLUE) (e.g.Veldkamp and Fresco 1996; Verburg et al. 1999; Overmars et al. 2007; Verburg and Overmars 2009). The CLUE modelling framework proves that it is possible to model LUC on a more detailed level, taking into account driving forces at different spatial levels. However, as the CLUE modelling approach is based on the competition between land use functions, it suggests some form of top-down land use planning. However, LUC is not always policy driven and is in less developed countries often related to local mechanisms. Moreover, CLUE does not consider the effects of the uncertainties in the input data on the results of LUC modelling. The objective of this study is to develop a new LUC model to assess the development in land availability for bioenergy crops on a detailed spatial level, taking into account the dynamics of several other land use functions and the uncertainties in drivers of LUC. The model is specifically developed for less developed countries characterised by subsistence farming, a low density of infrastructure, and a lack of top-down land use planning. The land use changes in these types of countries are driven by environmental and socioeconomic factors and are influenced by national or regional land use planning and policies to a much lesser extent. A multitude of driving forces and suitability factors are included in the model. The detailed spatial level, the number of dynamic land uses, the diversity in driving forces and suitability factors, and the possibility to model uncertainties in a spatially explicit way serves as a step forward in LUC modelling for less developed countries. This model is especially developed to assess the land availability for bioenergy crops and therefore provides opportunities to assess how iLUC effects are to be avoided. The technical characteristics of the model are described in Verstegen et al. . This paper will show the functionality of the model and methodological issues related to LUC modelling with a case study on the development of land availability for bioenergy crops in Mozambique towards 2030. Mozambique was selected as a case study area as it is a promising region for biomass production within southern Africa as a result of the availability of land (Batidzirai et al. 2006; Namburete 2006), the favourable environmental conditions for agricultural production (INE 2003; Batidzirai et al. 2006), and the current low agricultural productivity which offers a great potential for improvement. The main incentives for the government of Mozambique to focus on the development of a bioenergy industry are to decrease the country’s dependence on oil imports and to enhance energy security and socio-economic and sustainable development, especially in rural areas (Conselho de Ministros da república de Moçambique 2009). The fragmentation of current land use and the high spatial variation in climate and soil conditions in

124

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

Mozambique requires a high spatial resolution of land use change and bioenergy potential modelling. The methodology section will elaborate on the steps required to model land use change over time in a spatially explicit way. The Appendix provides all input data used for the land use modelling of Mozambique. The results section examines the typical results that can be produced with the model, including the uncertainties in output calculated from uncertainties in input and model parameters. The discussion and conclusion section will elaborate on the relevant outcomes of the Mozambique case study, and the consequences of the characteristics of the model and of the methodological choices made.

4.2 Methodology It is of key interest to assess how competition for land and related effects of iLUC can be avoided; therefore, the modelling of the land availability for energy crop production needs to take into account the land required for other land use functions. Land use requirements for crop and livestock production depend on the developments in food demand and agricultural productivity. Consequently, land use is dynamic over time. This study includes the demand for food, feed and materials (including wood) which results in a claim on land for crop production and grazing area as well as in deforestation. In order to project the dynamics in these land use functions over time, future developments regarding the main drivers for LUC need to be identified and quantified.

4.2.1 Drivers of land use change The main LUC drivers are the developments in the demand for food, feed and materials and the productivity of the agricultural sector. The demand for domestically produced food and feed is related to developments in population size, GDP, food intake per capita and self-sufficiency ratio (SSR, i.e. the extent to which domestic supply meets domestic demand) (FAO 2003b). The amount of land required to meet the total demand for food, animal products and materials depends on the efficiency of the agricultural sector. Developments in the efficiency of crop production are related both to the exploitable yield gap, i.e. the gap between current yields and agro-ecological or maximum attainable yields (FAO 2003b), and to the rate of technology adoption, i.e. the implementation pace of improvements in crop production. The efficiency of livestock production is related to the distribution of supply between types of production system (pastoral or mixed), the feed composition (the share of feed supplied by grazing, scavenging, residues and feed crops), and the feed conversion efficiency (the amount of animal product per unit feed) of the

125

SHADES OF GREEN

production systems. The land requirements for feed crops and pastures depend on the feed crop yield and the carrying capacity of pastures. The development in drivers of demand and the development in agricultural productivity in Mozambique are described in Table 1 and are quantified in more detail in the Appendix (Section 4.6 Table 4.2 and Table 4.3). Developments in the demand for wood are related to the developments in total population, the ratio between urban and rural population, the adoption of improved cooking technologies, the domestic use of poles and other timber, and the export quantity of industrial round wood. The domestic wood supply can be roughly divided into two categories: wood that is sustainably extracted from the forest and wood whose logging results in deforestation. As this study focuses on LUC dynamics, only the wood demand that leads to deforestation has been included, defined as the illegal and unsustainable wood harvesting in forest areas. Hence, sustainable logging and logging in other woodland is not included. The developments in wood demand and deforestation in Mozambique are quantified in the Appendix (Section 4.6 Table 4.4).

4.2.2 Scenario approach Since it is uncertain how LUC drivers evolve and the prediction of land use developments is problematic (Verburg et al. 2004), a scenario approach was used to explore potential long-term developments in LUC driving forces. The use of scenarios to explore potential LUC developments has already been demonstrated by Stengers et al. (2004), Westhoek et al. (2006), De Vries et al. (2007) and Hoogwijk et al. (2005; 2009). In this study, the narratives developed for the Special Report on Emissions Scenarios (SRES) of the Intergovernmental Panel on Climate Change (IPCC) (Nakicenovic et al. 2000) were translated into specific scenarios for Mozambique to develop a consistent set of assumptions for the assessment of future land use dynamics. Two divergent storylines were developed: a Business-As-Usual (BAU) scenario based on the B2 storyline and a Progressive scenario based on the A1 storyline. The development in the key drivers of LUC, demographics, consumption patterns, and GDP, are rather unpredictable, which justifies the creation of divergent scenarios regarding the development of these driving forces. Still, for reasons of transparency, the developments in these main drivers are kept equal for the two scenarios and an uncertainty analysis in which these drivers are modelled stochastically is assumed to be the most suitable way to address the sensitivity of the results to these parameters. This implies that population, GDP, diet and SSR will change over time, but that the rate of change is equal for the two scenarios. The divergent storylines were used to explore possible developments in technological, institutional and societal changes which result in changes in productivity in the agricultural sector. Table 4.1 presents a qualitative description of the current status and the characteristics of the 126

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

scenarios. The development in the drivers is quantified for the two scenarios in the Appendix (Section 4.6: Table 4.2, Table 4.3, Table 4.4). Table 4.1: Current status and development of key drivers of land use change for the Business-AsUsual and Progressive scenario. Parameter

Current

BAU Scenario

Population a

Population size: 22.9 million people Average population density: 29 p/km2 Rural population: 63.5%

Population size: 33.9 million people by 2030. Average population density: 43 p/km2 in 2030 Rural population: 46% in 2030

GDP b

Average annual growth rate: 8% (19942007)

Average annual growth rate: 6.6% up to 2030.

Diet c

Average caloric intake: 2050 Kcal/cap/day. Animal product consumption: 1.4% of caloric intake. The average diet consists mainly of roots and tubers, and to a lesser extent of cereals.

Average caloric intake: 2550 Kcal/cap/day (2030) Animal product consumption: 3.7% of caloric intake. The proportion of the total diet supplied by cereals increases at the cost of roots and tubers.

SSR

Self sufficient for most food crops. Main import products: cereals and animal products. Main export products: tobacco, cotton, sugar and wood.

The SSR ratios will remain constant up to 2030 and export will remain at today’s levels.

Farming practices

Farming system: subsistence farmers (95%) Cultivation area size of 0.5-1.4 ha in shifting cultivation, clearing by burning

Continuation of current practices, a modest shift towards commercial farming.

Technological adoption d

Low adoption of improved seeds (5-10% of farmers), fertilizers (3.9%), pesticides (4.5%), irrigation (5%) and animal or mechanical traction (11% and 3%).

Continuation of current Increasing share of trends in input levels. farmers have access to improved seeds, fertilizers, agrochemicals, knowledge, machinery and irrigation.

Agricultural productivity e

Very low agricultural productivity (e.g. average maize yields are <1 ton ha-1) and cropping intensity (60%).

A modest increase in yield (0.6% p.a.) and cropping intensity (0.5% p.a.), in line with historical trends.

Higher agricultural productivity due to higher yields (3.5% p.a.) and increased cropping intensity (2% p.a.).

Livestock sector f

Low livestock numbers. Development of the sector is hampered by the lack of disease control. Cattle and ruminants are kept in

No additional policies to prevent diseases. Partial shift from

Effective policies focusing on disease control. Strong shift

127

Progressive Scenario

Shift towards commercial farming, shifting cultivation is progressively abandoned.

SHADES OF GREEN

Deforestation g

Bioenergy implementati on

pastoral systems and extensive mixed systems. Low feed conversion efficiencies in both systems.

pastoral systems to mixed systems. Modest growth in feed conversion efficiency.

towards mixed systems. Increased feed conversion efficiencies in both pastoral and mixed systems.

High deforestation rates due to high wood demands and lack of enforcement.

No additional policies, regulation and enforcement. Continuation of current trends in deforestation.

Additional policies, regulation and enforcement to prevent further deforestation

Bioenergy projects implemented in a developing institutional and regulatory framework.

No major changes in policies, technologies and current practices.

Bioenergy is implemented in a controlled and sustainable environment.

a

Population figures are based on FAO (2010a) and UNDP (2008). Population density varies strongly between regions. b GDP figures are based on IMF (2007; 2010). Despite the high economic growth, about 54% of the population remains in poverty today. The poverty line is set based on the value of a basket of basic needs goods based on consumption patterns of the poor and varies by province (Fox et al. 2005). c Figures on dietary intake are based on FAO (2003b; 2010a). Due to an unequal distribution of food and to the dietary composition, a large proportion of the population suffers from undernourishment. d Figures on the use of inputs are derived from Bias and Donovan (2003), INE (2003), and the World Bank (2006). The lack of both financial resources and markets cause the low adoption of inputs (World Bank 2006). e Figures are based on INE (INE 2003)and FAO (FAO 2010a).There is little incentive to increase production because of a lack of markets, insufficient access to markets, a lack of a viable market prices and insufficient storage capacity. The cropping intensity is the area harvested expressed as a percentage of the arable area (FAO 2003a). Arable land includes temporarily (less than 5 years) fallow land, but abandoned land resulting from shifting cultivation is not included in this category (FAO 2005c). f Livestock figures are based on (Otte and Chilonda 2002; INE 2003; FAO 2005d; FAO and WFP 2010). Pastoral systems are based on extensive grazing, whereas mixed systems are based on both grazing and feed crops. There is a strong spatial variation in the distribution of livestock: cattle is mainly concentrated in the south, due to the prevalence of the Tsetse fly and related diseases in central and northern parts of Mozambique (Timberlake and Reddy 1986; INE 2003; Maposse et al. 2003). Pigs, chickens and small ruminants are kept all over the country with higher concentrations in the central and northern parts of Mozambique. g Estimations of deforestation rates in Mozambique vary between 0.2 en 5.6% (Mangue 2000; Pereira et al. 2001; Del Gatto 2003; Nhancale et al. 2009; FAO 2010b). Causes of deforestation are high demands for fuel wood and charcoal (Cuvilas et al. 2010), high export rates of industrial round wood (80%)(FAO 2010b), poor control on temporarily logging concessions and illegal logging (Del Gatto 2003; MacKenzie 2006; Nhancale et al. 2009; Cuvilas et al. 2010).

4.2.3 Land use modelling Due to variations in agro-ecological conditions, the yields of crops, pasture and wood are spatially highly heterogeneous. Therefore, the total amount of land required to meet the demand for food, wood and animal products is directly related to the location of the 128

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

specific land use class. Several studies on LUC have developed methodologies for land use allocation. Land use classes can be modelled dynamically (related to LUC drivers), passively dynamically (not linked to a demand but susceptible to change when other land use functions expand), or statically (excluded from any LUC). The characteristics of the land use classes that are modelled dynamically, passively dynamically and statically in this study are described in Section 4.6.2 and Table 4.5 of the Appendix. In this study, the allocation of land to dynamic land uses classes is based on the suitability of the location for a specific land use class which is defined by a combination of several selected spatially explicit suitability factors. In some LUC models, a multiple regression model is applied to identify the driving forces and assess the influence of these factors on land use, e.g. in Veldkamp et al. (1996) and Verburg et al. (1999). However, these models require historical data of the land use patterns, and these are not available for Mozambique. Moreover, extrapolation of regression analysis may produce dubious results as historical driving forces for LUC may no longer be detected or no longer be relevant. This is especially true for countries where historical developments are characterised by discontinuities, such as war or natural disasters. The limited predictive value and uncertainty of the causality are drawbacks of the statistical quantification based on historical developments (Veldkamp and Verburg 2004). In this study, driving forces have been identified by expert consultation and literature review. In addition to environmental driving forces, land use modifications are strongly influenced by socio-economic and policy-related issues (Lambin et al. 2001). However, incorporating such factors is hampered by a lack of spatially explicit socio-economic data and the methodological difficulties in linking socio-economic and environmental data (Veldkamp and Lambin 2001). In this study, both environmental and socio-economic drivers are taken into account. In line with Veldkamp et al. (2001), proxy indicators have been selected to represent socio-economic driving forces. The suitability factors selected in this study are: proximity to same land use class, distance to roads, distance to water, distance to cities, distance to forest edge, potential yield level, population density, cattle density and conversion elasticity. For each land use class, a suitability map was constructed based on the spatially weighted summation of a specific set of individual suitability factors. For each suitability factor, the direction of the relation (e.g. does the suitability increase or decrease with distance to road), the type of correlation (exponential, linear, inversely related), and the maximum distance of effect (e.g. up to what distance from the road does the road still influence LUC) were determined. The characteristic of the suitability factors for land use allocation are further explained in the Appendix (Section 4.6.2, Table 4.6). Areas that are not suitable (e.g. steep slopes) or not allowed (e.g. conservation areas) to be converted to agricultural land were excluded. Table 4.7 of the Appendix provides an 129

SHADES OF GREEN

overview of the excluded areas and the spatial data used. In some land use models (such as CLUE), the allocation of land has been based on the highest suitability of one land use class compared to the other land use classes (Verburg and Overmars 2009). This approach serves top-down land use planning, which regulates land use in such a way that land is used for the best possible application. Yet, in this study, a fixed order for allocation is used. The order of allocation applied in this study is (1) forest (deforestation does not compete with other land use functions) (2) cropland, (3) mosaic cropland-pasture, (4) mosaic cropland-grassland, and (5) pasture. This implies that the land use category “deforestation” is allocated first to the best suitable places for deforestation until the demand (for wood that results in deforestation) is met. Deforestation can only occur on land currently in use as forest. Subsequently, cropland is allocated to the locations best suitable for cropland, until the demand of that particular year is met by the production (Area x Location specific productivity x Management level). Subsequently, mosaic cropland-pasture, mosaic cropland-grassland and pastures are allocated. This order of allocation is based on the assumed economic importance and the labour requirements of these land use classes from the individual farmer’s perspective. This allocation mechanism is considered to be more realistic for less developed countries, where there is usually no top-down land use planning but a more locally driven LUC. Land is allocated to a land use class in time steps of one year. This allocation of land within one time step stops when the production of that land use class has met the demand for that particular time step. The amount of land required to meet the demand depends on the productivity of the land allocated, and on the agricultural efficiency during that time step. Once the land has been allocated, it cannot change to another land use class during the same time step (because in that case supply will not meet demand). The total allocation is completed for one time step when all the land use classes have been allocated and the production of these land use classes meets the total demand for that particular time step. This results in a new land use map for that time step. The modelling comprises a feedback loop: the land use resulting from the allocation in time step t serves as input for the allocation in time step t+1. Figure 4.1 shows how the dynamics of land use classes are modelled and how they influence the land availability for energy crop production. LUC drivers (population, diet, GDP and SSR) determine the demand for food crops and animal products for each time step (Section 4.2.1). The scenario characteristics determine the developments in the productivity of both crop cultivation and the livestock sector, the fuel wood demand per capita, and the deforestation rates (Section 4.2.2). Based on the demand for animal products and the efficiency of the livestock sector, the amount of feed crops and pasture is calculated. Based on population growth and the fuel wood demand per capita, the total 130

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

fuel wood demand is calculated. Based on a specific set of suitability factors, the excluded land, and the order of allocation, land is allocated to the different land use functions (Section 4.2.3). This results in a new land use map. Based on this land use map and a map of the areas that are excluded for bioenergy crops (such as community land) in addition to the areas already excluded for LUC, the land availability for bioenergy crops is determined. The land use map which results from the land use allocation of year t serves as input for the land use allocation in year t+1.

Figure 4.1: Overview of the modelling of land availability for bioenergy crops.

In order to enable the modelling of future land use as depicted in Figure 1, a spatiotemporal land use model has been developed based on the building blocks of the PCRaster Phyton framework (Karssenberg et al. 2010; PCRaster 2010). The key inputs for the PCRaster Land Use Change model (PLUC) are: time series of demand and productivity development, dynamic land use classes, suitability factors per land use class, the initial land use map that designates the initial configuration of these land use classes and several 131

SHADES OF GREEN

maps of suitability factors (e.g. population density and distance to road). The parameterisation of these and additional inputs are discussed in the Appendix. The major advantage of this model framework is its ability to deal with stochastic input data. This enables spatio-temporal Monte Carlo (MC) runs that evaluate uncertainty propagation. PLUC can stochastically model time series (e.g. crop demand and agricultural productivity), spatial input parameters (e.g. population density and productivity), and characteristics of suitability factors (e.g. the maximum distance of effect in the distance to road). The stochastic inputs can be based on different error models: an uniform distribution between two values, a normal distribution given the mean and fixed standard deviation (SD), and a relative distribution given the mean and a relative SD. When a uniform error model is applied, all values between the upper and lower limit have equal probability. The normal error model has a normal distribution of probabilities, with 95% of all selected values within the range of the mean + /- 1.96 SD. The relative error model also has a normal distribution, but with the SD relative to the mean. The probability distribution of stochastic inputs is equal for each time step. More information on the stochastic input variables and the applied error models is provided in the Appendix Section 4.6.2 and Table 4.8. The probability of the availability of land for bioenergy can be calculated by means of an MC analysis. The probability can not only be analysed at a grid cell level but also at a provincial or national level. More information on the technical characteristics and stochastic input modelling of PLUC can be found in Verstegen et al. (2011). The software package Aguila enables the visualisation of the results of the PLUC model for every individual time step (de Jong 2009; Karssenberg et al. 2010). It can show the development in LUC and the land availability for bioenergy crops for a deterministic run, as well as the development in the probability of land availability for bioenergy crops for a MC run.

4.3 Results This section provides some selected results and uncertainty analyses that are illustrative for the possibilities of the developed model. A more elaborate overview of typical results and uncertainty analysis can be found in Verstegen et al. (2011).

4.3.1 Developments in demand, productivity and land requirements Figure 4.2 depicts the total domestic food and non-food production for the timeframe 2005-2030. The total food demand that needs to be produced domestically is expected to increase from 11.5 Mt in 2005 to 24.7 Mt in 2030. In the BAU scenario, the total wood 3 3 demand increases from 19.4 million m in 2005 to 39.0 million m in 2030, of which 43% is

132

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

expected to result in deforestation. In the Progressive scenario, the total wood demand 3 increases to 20.2 million m for 2030. This lower wood demand results from the adoption of improved stoves and alternative fuels. ’As in this scenario deforestation is to be 3 prevented, 9.2 million m should be produced in alternative ways by 2030. The prevention of deforestation is a result of strong policy measures assumed in the progressive scenario.

Figure 4.2: Total food and non-food crop demand in timeframe 2005-2030 considering the developments in population growth, dietary intake and SSR ratios. The error bars indicate the range in demand given the lower and higher projections for population growth (32 million - 36 million people in 2030) (UNDP 2008) and dietary intake (2050 - 2980 Kcal/cap/day in 2030) (FAO 2003b).

Figure 4.3: Development in crop and livestock productivity in the BAU and Progressive scenarios in the time frame 2005-2030, normalised for the productivity levels of 2005 (2005=1). The bandwidths represent the range of the uniform distribution of the stochastic input of yield developments for the BAU and Progressive scenarios.

133

SHADES OF GREEN

The land required to meet the demand depends on the developments in agricultural productivity. In Figure 4.3, the developments in the productivity of crop cultivation and livestock production are presented for the two scenarios. It shows the normalised productivity increase compared to the level of the year 2005, based on the weighted summation of the productivity increase per crop (based on the proportion of cultivated area) and the weighted summation of the productivity increase per animal product (based on the proportion of total volume). The bandwidth of the curves of the development in crop productivity in the BAU and Progressive scenarios represent the range of the stochastic input of the maximum attainable yield (see Appendix, Section 4.6.2 on stochastic model inputs). Figure 4.4 presents the land requirements to meet the total crop and grazing demands for the BAU and the Progressive scenarios, assuming the same distribution of cropland and pasture over potential yield classes as in 2005. In the BAU scenario (left), there are two reasons why the land required for crops and pasture increases: an increased demand caused by population growth and a rise in food intake per capita, and a relatively low growth in productivity. In the Progressive scenario (right), both pastures and arable land areas decline due to increased yield levels of pasture and crops, and a greater efficiency in livestock production. The upper sections of the bars (grey shade) indicate the additional land required due to low cropping intensities, i.e. it accounts for the land that is left fallow for a short time. The error bars indicate the uncertainty in the total land requirements given the uncertainty in the development in demand (see Figure 4.2) and the uncertainty in crop productivity in both scenarios (see Figure 4.3) The positive error value is bigger than the negative error value as a consequence of the uncertainty distribution of demand (see Figure 4.2), which is also skewed. By 2030, the land requirements in the BAU scenario are 3.3-3.7 times higher than in the Progressive scenario. As the location of the land is an important determinant factor for the yield of crops and pastures, the actual land requirements can only be calculated when spatial land use allocation is taken into account. Therefore, the PLUC model has been applied to assess the land requirements in a spatially explicit manner.

134

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

Figure 4.4: Land requirements for livestock grazing and crop production for the timeframe 20052030 for the BAU (left) and Progressive (right) scenarios, given the same distribution over productivity classes of pasture and arable land as in 2005. The error bars represent the range in total land requirement given the uncertainties in total demand (Figure 4.2) and productivity (Figure 4.3).

4.3.2 Deterministic spatial modelling results Figure 4.5 displays the results of the deterministic simulations of LUC for the BAU and the Progressive scenarios. In the deterministic run, the uncertainties in input parameters have not been modelled. Figure 4.5 shows the land use for 2005 (time step 1, same for both scenarios), 2015 (time step 11), and 2030 (time step 26). In the subsequent maps for 2015 and 2030, it is apparent that cropland, mosaic cropland-pasture, mosaic croplandgrassland and pasture areas are expanding in the BAU scenario, whereas these land use types are contracting in the Progressive scenario. In the BAU scenario, the shift towards pure or mosaic cropland and pastures is most profound close to main cities and in proximity to the road network. The expansion of agricultural land use is mainly at the expense of forest (76%) and shrubland (21%). In the Progressive scenario there is a shift from the more extensive mosaic cropland towards specialised cropland close to the main cities and in proximity to the road network. Extensive mosaic cropland and pastures are progressively abandoned due to the intensification of crop and livestock production. This is most apparent in the remote semi-arid and less populated areas (Gaza and Tete provinces). Another important difference between the scenarios is the development in autonomous deforestation, in addition to forest converted to agricultural land uses. In the BAU scenario, deforestation is most apparent along the main road network. The location of deforestation is especially influenced by the ‘proximity to road infrastructure’ and ‘distance to forest edge’ as included in the allocation procedure. This is in line with the 135

SHADES OF GREEN

observed land use changes in Manica province where land use changes were most common within 5 km from the roadside and where deforestation was the most dominant LUC (Jansen et al. 2008). The expansion of the deforested areas is most profound in the first ten time steps. Due to the assumed regeneration of forest, the expansion of deforestation slows down in the BAU scenario. In the Progressive scenario, it is assumed that deforestation can be prevented from 2011 onwards, and the effects of deforestation are no longer visible after 2020. The developments in land availability for biofuels over the timeframe 2005-2030 are presented in Figure 4.6. The red areas indicate areas that are not available for bioenergy crops. These areas have been excluded because they are used for other land use functions, such as cropland, pasture, forest and urban areas, or because they are not suitable (e.g. regularly flooded areas or steep slopes). In the BAU scenario, the available land area decreases as land required for pasture and crops expands. As the expansion of cropland and pasture areas occurs mainly in the densely populated areas close to the main cities and road network, the land available for bioenergy is decreasing most rapidly in these areas (e.g. along the main North-South road and the Beira corridor). The areas which remain available for bioenergy crops are the more remote and less productive areas: in the central northern parts (Cabo Delgado, Niassa and Nampula provinces; mainly moderately to marginally productive), the north-western parts (Tete province; mainly marginally to non-productive) and south-western parts (Gaza province; mainly marginally to non-productive). In the Progressive scenario, the area required for crop and livestock production decreases over time. Mainly areas with an initially high proportion of mixed cropland-grazing become available. These areas are mostly situated in the South-East (Inhambane province; moderately to very productive), North-East (Nampula province; marginally to very productive) and North-West (Tete province; marginally to nonproductive). Figure 4.7 displays the development in available land for bioenergy crop production until 2030 according to the deterministic run for the two scenarios. For the BAU scenario, land availability decreases over time from 9.1 Mha to 7.7 Mha. For the Progressive scenario, the land availability for bioenergy crop production increases from 9.1 to 16.4 Mha. As the Progressive scenario assumes that no deforestation will take place, the wood demand needs to be supplied in alternative ways. In 2030, 0.4 Mha of forest plantation would be required to compensate for the fuel wood consumption (based on the assumed average productivity of the available land in 2030). In Figure 4.8, the developments regarding land availability for bioenergy crops for 2005, 2015 and 2030 for the two scenarios are depicted; the available land is differentiated for 136

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

five productivity classes. The productivity classes provide the percentage of the maximum attainable yield given the level of agricultural management. The non-productive category includes areas that produce <20% of the maximum attainable yield. Although the marginal suitable soils are not very productive (20-40%), these areas should not be excluded beforehand for energy crops (as productivity of the area is just one of many criteria for site selection). However, suitable bioenergy crops for these areas should be selected with care. The most productive areas available for bioenergy are located close to the Malawian border and in the central north-eastern part of the country (Zambezia and Nampula provinces).

137

SHADES OF GREEN

Figure 4.5 : Land use dynamics up to 2030 for the BAU (upper maps) and Progressive scenarios (bottom maps).

138

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

Figure 4.6 : Land availability for bioenergy crops in 2005, 2015 and 2030 for BAU and Progressive scenarios (based on deterministic runs). Red areas indicate the areas that are not available, whereas the green areas are the areas available for bioenergy crop production.

139

SHADES OF GREEN

Figure 4.7 The development of land availability for bioenergy crop production over time for the BAU (lower trend line) and Progressive scenarios (upper trend line).

Figure 4.8: The development of land availability over time differentiated for suitability classes for the BAU (left) and the Progressive scenarios (right).

4.3.3 Uncertainty in spatial modelling results The results described above were generated with deterministic runs of the model. By running an MC sample of 500 realisations, the effect of the uncertainties in input variables and model parameters can be assessed on the land availability for bioenergy crops. The uncertainties included in this MC analysis are listed in Table 4.8 of the Appendix. Figure 4.9 presents the uncertainty in land availability for bioenergy crops in the BAU scenario for the time frame 2005-2030. The blue- grey areas indicate the 95% confidence interval, which ranges from 6.1-8.2 Mha for 2030. This means that 95% of all realisations of the MC sample result in a land availability within this range. The median, represented by the black 140

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

line, is slightly different from the deterministic run of the BAU scenario (see Figure 7), because the deterministic input of demand is not equal to the mean of the uniform distribution of the stochastic input of demand (see Figure 4.2) and because the non-linear relationships between the uncertainty in input parameters and the uncertainty in the results.

Figure 4.9: Uncertainty in land availability for bioenergy crops in the BAU scenario for the time frame 2005-2030.

For all the realisations of the MC sample, the land availability for bioenergy crops has also been mapped spatially. The probability of the availability of the grid cell for bioenergy crops was determined by combining the maps of all these realisations of the MC sample. Figure 4.10 depicts the probability of land availability for bioenergy crops in the BAU scenario for three time steps. A value 0 indicates that the grid cell is unavailable in all realisations, whereas 1 indicates that the grid cell is available in all realisations. In 2005 (time step 1), there is already some uncertainty (the grid cells do not only have the values 0 and 1, but also values between 0 and 1). This is because the initial land use map was calibrated based on the yield map, whereas the yield map is a stochastic input in this uncertainty assessment in time step 1. The uncertainty increases over time because the uncertainties of the input parameters increase over time and because of the feedback loop in the model. As the rate and shape of expansion of the dynamic land uses differ for each realisation in the MC sample, the uncertainty is most apparent at the edges of expansion (see Figure 4.10). The areas that are excluded based on non-dynamic land use (e.g. nature conservation areas) do not contribute to the uncertainty and dominate the total excluded land area. Therefore, the uncertainty in land availability is restricted to certain areas.

141

SHADES OF GREEN

The MC assessment can be used to exclude areas for which it is uncertain whether they will be occupied by other land use functions. For example, only the land that is available in 95% of the realisations of the MC analysis could be labelled ‘available for bioenergy crops’. The application of this threshold results in a lower estimation of land available for bioenergy crops than the results of the deterministic run. For the BAU scenario, the deterministic run results in 9.1 Mha available in 2005, but when a threshold of 95% is applied in the MC run, only 9.0 Mha is available. For 2030, the deterministic run indicates a land availability of 7.7 Mha, whereas the 95% confidence interval of the MC analysis indicates 5.7 Mha. As uncertainty is propagated over time, the difference increases over time between the results of the deterministic run and the 95% confidence thresholds of the MC analysis. The amount of land labelled as available decreases when the confidence threshold is increased (e.g. to 99% confidence). In addition to the probability on grid cell level based on several stochastic inputs, the PLUC model enables other types of uncertainty analyses, such as the probability of land availability based on uncertainty in a single parameter, the probability of land availability at a provincial level, or the probability of exceeding the threshold of total biomass production given a specific bioenergy crop. More information on error propagation and uncertainty assessments on several spatial levels can be found in Verstegen et al. (2011).

142

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

Figure 4.10: Probability of land availability for bioenergy crops at grid cell level for several time steps for the BAU scenario based on stochastic input variables of several parameters (see Table 4.8).

143

SHADES OF GREEN

4.4 Discussion and conclusion This study has determined the potential development of spatially explicit land availability for bioenergy crop production, taking into account the dynamics in agricultural land use and deforestation. As bioenergy crop production was not included as a dynamic land use class, the competition for land or indirect land use change as a result of bioenergy crop production has not been modelled in this study. If competition between bioenergy crops and other land uses is to be modelled, the implementation of bioenergy crops should relate to a projected demand (e.g. national biofuel blending targets), as demonstrated for Brazil by Lapola et al. (2010). However, the starting point of this study is that competition for land is to be avoided, and therefore land was excluded if it was (potentially) in use for other land use functions. In this study, a fixed allocation order was applied to the other land use classes; consequently, the competition between the land use functions based on macro-economic parameters was not modelled. In order to properly model competition between land uses, extensive information is required on market developments, price elasticity and policies. The interaction between macro-economic and land use modelling is beyond the scope of this study. Land required for food production has been excluded from the conversion to bioenergy crop production. As the estimated average food intake by 2030 (2550 kcal/cap/day) is still below the estimated average of developing countries in 2030 (2980 kcal/cap/day) (FAO 2003b), it can be argued that for food security reasons the domestic production and the related claim of land should be higher. On the other hand, food security is not only an issue of food availability but also of food accessibility. It is assumed that the current SSRs of food continue towards 2030. However, it may be argued that the SSR ratios will decrease, in line with the FAO projection for sub-Saharan countries (Otte and Chilonda 2002; FAO 2003b). This would imply a lower domestic production and therefore a reduced land claim. The returns from biofuel export and/or the decreased expenditure on oil imports could contribute to food security by increasing purchasing power. Only two of the nine suitability factors for land use allocation (‘number of neighbouring cells in the same land use class’ and ‘conversion elasticity matrix’) change over time as a result of the output of the model (output of t is input for t+1). However, other suitability factors such as ‘population density’, ‘distance to road’ and ‘distance to cities’ are expected to change over time due to a shift from rural to urban population, the development of new infrastructure, and the emergence of new economic centres. This will affect future LUC and will therefore influence the location and the amount of land available for bioenergy crops. However, the dynamics of these suitability factors were not modelled

144

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

over time in a spatially explicit manner because of a lack of knowledge and data regarding these future developments. The calibration of the ESA 2005 land use map with statistics was used as a starting point for the land use modelling. The ESA 2005 land cover map was the most recent detailed map available. In addition, three-year averages of production statistics were used to correct for weather-related annual fluctuations, and statistics were only available up to 2008. Validation of high spatial resolution is difficult and sometimes even impossible when validating future land use (Verburg et al. 1999). Historical analyses could contribute to the validation of the model, but in the case of Mozambique this was not realistic because of the lack of historical (spatial) data and discontinuities in historical land use developments due to the struggle for independence and the civil war. Therefore, it was not possible to assess whether modelled patterns in LUC match historical and current LUC dynamics. With more spatially detailed datasets of land use becoming available, future validation of the model will be possible. For the model developed in this study and the developments in detailed spatial land use modelling in general, there is an increased requirement for high quality and more detailed spatial data. This was also concluded by Schmit et al. (2006) for assessments in Europe. However, it is especially true for less developed countries such as Mozambique, which are characterised by dynamic developments and where systematic (statistical) data gathering is still in a developing phase. As the scenarios developed in this study are highly divergent, the outcomes of the assessment provide a wide range of possible future developments. The land available for bioenergy crops in the BAU scenario decreases over time from 9.1 Mha to 7.7 Mha. This is caused by the expansion of cropland and pastures required to meet the increasing demand for food and animal products. The areas which will remain available for bioenergy crops are the more remote and less productive areas: in the central northern parts (Cabo Delgado, Niassa and Nampula provinces; mainly moderately to marginally productive), the north-western parts (Tete province; marginally and non-productive) and south-western parts (Gaza province; marginally and non-productive). In contrast, the Progressive scenario shows increasing land availability for bioenergy crops: from 9.1 Mha in 2005 to 16.4 Mha in 2030. In this scenario, mainly areas with an initially high proportion of mixed cropland-grazing will become available, and this is caused by more efficient agricultural production. These areas are mainly situated in the South-East (Inhambane province; marginally to very productive), North-East (Nampula province) and North-West (Tete province). By 2030, the land available in the Progressive scenario is almost twice as high as in the BAU scenario. This estimated potentially available land for bioenergy crops in the 145

SHADES OF GREEN

timeframe 2005-2015 is in the same order of magnitude as the range found in other studies, namely 6.0-13.2 Mha (Batidzirai et al. 2006; DNTF et al. 2008; Econergy 2008). However, in this paper, the potential is determined in a spatially and temporal explicit and probabilistic way. In the progressive scenario it assumed that deforestation is prevented from 2011 on. This implies that, in this scenario, 0.4 Mha of forest plantation is required in 2030 to meet the wood demand that was alternatively fulfilled by illegal logging. However as current developments show that deforestation is still continuing, a total halt of deforestation by 2011 is clearly unrealistic. The assumptions made in the progressive scenario are only to illustrate how land use dynamics would look like if deforestation would have been prevented from 2011 on. The 95% confidence interval for the amount of available land for biofuels in the BAU scenario in 2030 ranges from 6.1-8.2 Mha. The spatial assessment of the MC analysis of land availability shows the probability of land availability on a grid cell level. This demonstrates that especially on the edges of the expansion of land use classes, the availability of land is uncertain. If a threshold of 95% certainty is applied, i.e. only land that is available for bioenergy in 95% of the realisations, 26% less land is labelled as available for biofuels in 2030 compared to the deterministic runs. The significant effect of the uncertainty in input parameters on the modelled land availability demonstrates the importance of the spatio-temporal modelling of uncertainty. However, only the uncertainty in some of the model input and parameters was included in the MC analysis (time series demand and productivity; spatial maps of productivity, population density, cattle density and elevation; window length and maximum distance of effect of the suitability factors), while the impact of uncertainties in other inputs and parameters could also be significant (such as the efficiency of livestock sector, proportion of the demand fulfilled by each land use class, the weighting of the suitability factors and the conversion elasticity matrix). By modelling the uncertainty in single parameters, it would be possible to further assess the relative contribution to the uncertainty in the amount and location of land available for bioenergy crops. Additional research into the uncertainties in land use modelling is therefore required. It can be concluded that in both scenarios a considerable amount of land may become available for energy crops, which does not conflict with food production. However, the decreasing potential in the BAU scenario indicates a higher competition for land in the future, which could hamper the development of a sustainable large-scale bioenergy sector in Mozambique. Therefore, it must be stressed that a large-scale sustainable bioenergy sector can only be established if it is developed simultaneously with a more productive and sustainable agricultural sector. This implies a discontinuation of current trends: a shift away from subsistence towards commercial farming and from pastoral towards mixed 146

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique 1

livestock systems. This should result in an annual yield increase of 3.5% and an increase in livestock efficiency of 2.5% per annum. This requires changes in agricultural management (especially deployment of fertilizer and improved seeds), development of regional or national markets, improved logistics, training and better overall capacities and governance of the agricultural sector. However, it is questionable if, and within what timeframe, the required conditions for such a transition could be met in Mozambique to realise the outcomes of the progressive scenario. The land availability for bioenergy crops modelled in this study is the land available when ILUC is to be prevented. In this approach, bioenergy production may not displace agricultural land. However in practice, bioenergy production will compete with other land use functions. It is likely that bioenergy producers will look for best suitable locations based on agro-ecological suitability and accessibility, which will (partly) be the locations also best suitable for current agricultural practices. Therefore, other land (than indicated as available in this study) could be used for bioenergy crops, but this will most likely result in indirect land use changes. To what extent this displacement will occur, depends on the corporate social responsibility of the company, policies and regulations of the Mozambican government and the criteria set for international trade of bioenergy. The land use model developed in this study is an advanced tool to assess future dynamic land use and land availability for bioenergy crops. The bottom-up approach, the number of dynamic land uses, the diverse portfolio of LUC drivers and suitability factors, and the possibility to model uncertainty is a step forward in modelling the land availability for bioenergy potentials. Spatially explicit assessment of land availability for bioenergy crops is an important precondition to assess potential bioenergy production, to design and implement bioenergy supply chains and logistics, and to assess the environmental and socio-economic impact of bioenergy production. The model has now been tailored to and demonstrated for Mozambique. Still, it is a flexible model which can be used for other countries when input data, rules and characteristics of suitability factors are adapted.

4.5 Acknowledgements This study is part of the Climate Changes Spatial Planning Programme and has been funded by the Dutch government, the European Commission and Shell. In addition, it has been partly funded by the Biorenewable Resources Platform, SASOL and UNEP. The authors gratefully acknowledge the following people, governmental and nongovernmental institutions, and agricultural, forestry and biofuel companies for their contribution to the data gathering and for sharing their knowledge and expertise: 1

For maize this means an increase of 0.9 ton/ha in 2005 to 4.5 ton/ha in 2030. 147

SHADES OF GREEN

CAPAGRI, DNER, DNTF, IIAM, ANE, CFM, Marc Schut, Anna Lerner, Célia Jordão, Peter de Koning, Helen Watson and many others. In addition, the authors would like to thank Professor Johan Sanders for his contribution to this study, Janske van Eijck for her contribution to the field work; Birka Wicke for commenting on the draft paper; Maarten Zeylmans van Emmichoven for his help with GIS and for bringing the team of authors together; and Kees Kwant, Marieke Harteveld and the members of the Biorenewable Resources Platform for their reflections on this study.

148

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

4.6 Appendix: Input data and model rules case study Mozambique 4.6.1 Land use change drivers in Mozambique The demand for domestically produced food and feed is mainly driven by changes in population size, GDP, food intake per capita, and SSR. The developments of these drivers are equal for the BAU and the Progressive scenarios and are depicted in Table 1 of the main document. The FAO outlook on crop production in Mozambique up to 2030 (FAO 2003b) was updated with the help of a more recent projection for GDP and population growth, historical trends of crop production and consumption based on time series of 1992-2008 (FAO 2010a), and figures provided by the national agricultural census (INE 2002; INE 2003). In the constructed time series of demand, 25 crops and 6 animal 2 products were distinguished. The current crop productivity is very low and the yield gaps are large in Mozambique. Although there is much room for improvement, historical trends show that increases in agricultural productivity are hard to accomplish. The FAO outlook (2003b) on crop production in Mozambique up to 2030, the national census (INE 2003), and the historical trends on productivity of 1992-2007 (FAO 2010a) were used to estimate future productivity in the BAU scenario. For the Progressive scenario, projected productivity figures were based on a wide range of literature sources and personal communication during field work throughout the country. In Table 4.2, the productivity increases per crop type are depicted for each scenario. The cropping intensity (area harvested/arable area) is currently low (60%) due to unsustainable land use (mineral exhaustion) and related long fallow periods as well as the seasonal rainfall pattern and the absence of irrigation. In the BAU scenario, the cropping intensity slowly increases to 66% for 2030 (FAO 2003b), whereas the Progressive scenario forecasts an increase to 100%.

2

Crops included in demand and productivity projections are: wheat, rice, millet, maize, sorghum, potatoes, sweet potatoes, cassava, other roots and tubers, sugar cane, pulses, vegetables, bananas, citrus fruits, other fruits, oil crops, ground nuts, sunflower, sesame, coconut, coffee, tea, tobacco, cotton, fibre. Animal products included in demand and productivity projections are: beef, mutton, pork, poultry, milk and eggs. 149

SHADES OF GREEN

Table 4.2: Annual yield increase in % per annum for main crops and yield levels in ton/ha of typical crop types in time frame 2005-2015 and 2015-2030 for Business-As-Usual and Progressive scenarios. BAU Crop type Cereals (wheat, rice, maize)a

PROGRESSIVE

BAU PROG

’05-‘15 ’15-‘30 ’05-‘15 ’15-‘30 Unit

Cereals (millet, sorghum)b

Example

‘05 ‘30

‘30

0.02

0.01

0.07

0.07

% y-1

Maize

0.9 1.1

4.5

Unit Ton ha-

0.02

0.01

0.05

0.05

% y-1

Sorghum

0.5 0.7

1.5

Ton ha-

-1

6.5 7.5

15

Ton ha-

17

60

Ton ha-

c

Roots and tubers (potato, cassava) 0.00 Sugar caned 0.02 Pulses and beanse 0.02 f

1 1

0.00

0.04

0.04

%y

Cassava

0.01

0.06

0.06

% y-1

Sugar cane 13

0.01

0.05

0.05

% y-1

Pulses

0.5 0.7

1.5

Ton ha-

1 1 1

Fruits, vegetables and oil crops

0.03

0.01

0.03

0.03

% y-1

Vegetables 5.5 7.6

9.0

Ton ha-

Tea and coffeeg

0.05

0.02

0.06

0.06

% y-1

Tea

1.9 2.5

2.9

Ton ha-

Cash crops (nuts, sunflower)h

0.02

0.01

0.04

0.04

% y-1

Sunflower 0.5 0.7

1.5

Ton ha-

Pasture

0.00

0.00

0.03

0.03

% y-1

Grass

3.4

Ton ha-

1.9 1.9

1 1 1 1

Totali 0.01 0.01 0.04 0.04 % y-1 Normalised 1.0 1.2 2.7 a Based on (Uaiene 2004; Coughlin 2006; World Bank 2006; Zavale et al. 2006; Republic of Mozambique 2010) and personal communication S. Kolijn. b Based on (Uaiene 2004; Coughlin 2006; World Bank 2006; Republic of Mozambique 2010). c Includes also sweet potato and other roots and tubers. Yield levels are based on yield levels in neighbouring countries (FAO 2010a) and personal communication S. Kolijn. d Based on commercial attainable yield in Mozambique (personal communication Hartmann, Maffambisse) and (Jelsma 2010). Current yield levels are very low: 13-16 ton ha-1 according to the FAO (2010a). However, reported yields of commercially grown sugar cane are 60-140 ton/ha (Jelsma 2010). e Based on (World Bank 2006; Republic of Mozambique 2010). f Based on historical and current yield levels in neighbouring countries (FAO 2010a). g Based on historical and current yield levels in Mozambique and neighbouring countries (FAO 2010a). h Also included coconut, tobacco and cotton. Based on historical and current yield levels in Mozambique and neighbouring countries (Magaia et al. 2005; Buss 2007; FAO 2010a; Republic of Mozambique 2010) and personal communication (S. Kolijn, R.L. Henriques, M. Barbosa). The exploitable yield gap is generally larger for subsistence crops than for industrial crops, as the latter is commonly cultivated in more commercial settings (Veldkamp and Fresco 1996). i The average annual yield increase for arable crops in total has been calculated by the weighted average of annual yield increases of individual crops related to the proportion of hectares used for the cultivation for each crop.

The livestock sector is characterized by a low productivity and is dominated by pastoral systems. In accordance with the FAO outlook (2003b), the BAU scenario assumes a modest shift from mainly pastoral to mixed systems for cattle, sheep and goats. For the Progressive scenario, the shift towards mixed systems is assumed to be more profound, so that the feed composition and feed conversion efficiencies in both pastoral and mixed systems approach the characteristics of livestock systems of Eastern African countries in 2030. The development of the efficiencies in the livestock sector are based on Bouwman et al. (2005), FAO (2003b) and Smeets et al. (2007), and are presented in Table 4.3.

150

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

Table 4.3: Livestock production system characteristics. Based on figures provided in (FAO 2003b; Bouwman et al. 2005; Smeets et al. 2007). P = Pastoral system M = Mixed system. Contribution of grass, residues, scavenging and feed crops to total feed intake are expressed as a fraction of total feed intake (=1). Current 2000 P M 0.95 0.05 0.97 0.04 0.00 1.00 0.00 1.00 0.70 0.30 0.00 1.00

BAU Scenario 2015 2030 P M P 0.89 0.11 0.82 0.90 0.10 0.84 0.00 1.00 0.00 0.00 1.00 0.00 0.65 0.35 0.61 0.00 1.00 0.00

M 0.18 0.16 1.00 1.00 0.39 1.00

Progressive scenario 2015 2030 P M P M 0.73 0.27 0.56 0.44 0.74 0.26 0.57 0.43 0.00 1.00 0.00 1.00 0.00 1.00 0.00 1.00 0.54 0.46 0.42 0.58 0.00 1.00 0.00 1.00

BEEF GOAT PORK POUL MILK EGGS Feed composition residues BEEF GOAT (fraction of total feed input) PORK POUL MILK EGGS Feed composition scavengingBEEF (fraction of total feed input) GOAT PORK POUL MILK EGGS Feed composition crops BEEF GOAT (fraction of total feed input) PORK POUL MILK EGGS

0.95 0.90 0.00 0.00 0.93 0.00 0.00 0.00 0.50 0.50 0.01 0.50 0.05 0.10 0.00 0.00 0.05 0.00 0.00 0.00 0.50 0.50 0.02 0.50

0.70 0.80 0.00 0.00 0.70 0.00 0.24 0.03 0.25 0.50 0.24 0.25 0.05 0.10 0.00 0.00 0.05 0.00 0.01 0.07 0.75 0.50 0.01 0.75

0.95 0.90 0.00 0.00 0.93 0.00 0.95 0.90 0.00 0.00 0.93 0.00 0.05 0.10 0.00 0.00 0.05 0.00 0.00 0.00 0.50 0.50 0.02 0.50

0.55 0.80 0.00 0.00 0.70 0.00 0.11 0.03 0.25 0.50 0.24 0.25 0.05 0.10 0.00 0.00 0.05 0.00 0.29 0.07 0.75 0.50 0.01 0.75

0.94 0.90 0.00 0.00 0.93 0.00 0.00 0.00 0.50 0.50 0.01 0.50 0.05 0.10 0.00 0.00 0.05 0.00 0.01 0.00 0.50 0.50 0.02 0.50

Feed conversion efficiency (kg feed/ kg animal product)

131 164 6.63 4.14 5.13 4.14

113 102 82 72 51 148 146 131 128 115 6.63 6.59 6.59 6.54 6.54 4.14 4.02 4.02 3.90 3.90 3.21 4.83 3.03 4.52 2.86 4.14 4.02 4.02 3.90 3.90

Animal production system characteristics Production system (fraction of total production)

Animal BEEF GOAT PORK POUL MILK EGGS

Feed composition grass (fraction of total feed input)

Max pasture (odt/ha) (%)

BEEF GOAT PORK POUL MILK EGGS

productivity GRASS

5.0

0.0%

151

0.63 0.80 0.00 0.00 0.70 0.00 0.17 0.03 0.25 0.50 0.24 0.25 0.05 0.10 0.00 0.00 0.05 0.00 0.15 0.07 0.75 0.50 0.01 0.75

0.95 0.90 0.00 0.00 0.93 0.00 0.00 0.00 0.50 0.50 0.01 0.50 0.05 0.10 0.00 0.00 0.05 0.00 0.00 0.00 0.50 0.50 0.02 0.50

0.0%

0.63 0.80 0.00 0.00 0.63 0.00 0.17 0.03 0.25 0.50 0.17 0.25 0.05 0.10 0.00 0.00 0.05 0.00 0.15 0.07 0.75 0.50 0.15 0.75

0.93 0.90 0.00 0.00 0.93 0.00 0.01 0.00 0.50 0.50 0.01 0.50 0.05 0.10 0.00 0.00 0.05 0.00 0.02 0.00 0.50 0.50 0.02 0.50

0.55 0.80 0.00 0.00 0.55 0.00 0.11 0.03 0.25 0.50 0.11 0.25 0.05 0.10 0.00 0.00 0.05 0.00 0.29 0.07 0.75 0.50 0.29 0.75

94 82 57 51 119 109 75 71 6.59 6.59 6.54 6.54 4.02 4.02 3.90 3.90 3.84 2.72 2.54 2.22 4.02 4.02 3.90 3.90 0.3%

0.3%

SHADES OF GREEN

Both scenarios assume a shift towards more efficient means of cooking and towards nonwood fuels, although this is more profound in the Progressive scenario. On the other hand, fuel wood requirements increase due to an increasing share of urban population (and the related shift from fuel wood to charcoal made from fuel wood), and due to increased consumption patterns. For both scenarios, it is assumed that the per capita demand and the amount of exported industrial round wood will remain at today’s level (2010). Because of the low densities and non-commercial character of fuel wood, the SSR ratio of fuel wood is 1. In the BAU scenario, the proportion of wood that is illegally harvested and results in deforestation is assumed to remain constant over time. For the progressive scenario it is assumed that in environmental and forestry policies will be enforced in such a way that no deforestation will take place after 2010. The area that is deforested as a result of illegal harvesting in forest areas depends on the productivity of the forest (standing stock) and the regeneration rate of forest. For both scenarios, it is assumed that the standing stock and regeneration rate will remain constant over time. Table 4.4 presents the wood demand and the proportion that results in deforestation. Table 4.4: Developments in wood demand and harvesting for two scenarios up to 2030. BAU

PROG

2000

2015

2030

2015

2030

Fuel wood use in rural areasa

Unit m3/ cap

0.8

1.0

1.0

0.8

0.5

Fuel wood use in urban areasa

m3/ cap

1.0

1.2

1.2

1.0

0.6

mln m3

13.1

13.8

14.8

13.8

14.8

Proportion round wood illegal harvestedc

%

50

50

50

0

0

Proportion fuel wood illegal harvestedc

%

95

95

95

0

0

Proportion round wood harvest in forestd

%

100

100

100

100

100

Proportion fuel wood harvest in forestd

%

50

50

50

50

50

Industrial round wood production

Harvest resulting in deforestation

b

d

%

90 90 90 0 0 years Regeneration ratee 10 10 10 10 10 a Based on figures provided by Pereira et al. (2001) and Cuvilas et al. (2010). b Based on figures of FAO (2010b). Currently, the lion share of industrial round wood is exported (80%). The total round wood production is assumed to remain at the same level. c Nhacale et al. (2009) and Del Gatto (2003) estimated that 50-70% of the total round wood production and 95% of the fuel wood is illegally harvested. d Fuel wood is mainly collected from dead wood, and charcoal production is primarily based on uncontrolled harvesting (Pereira et al. 2001; Cuvilas et al. 2010). It is assumed that half of the illegal fuel wood harvesting takes place in forests (and the remainder is collected or harvested in woodland/mixed land use types), whereas the illegal harvesting of round wood is fully allocated to forest areas only. As there are no incentives for illegal harvesters to harvest in a sustainable way, it is assumed that 90% of all illegal extraction of the wood directly results in deforestation. Informally or illegally removed wood, especially fuel wood, usually remains unrecorded; therefore, the actual amount of wood removal is difficult to estimate (FAO 2005b). e This is the estimated time in years required to regenerate the same amount of woody biomass as the original stock. Indigenous hard wood tree species need much longer to regenerate. However, for the use of fuel

152

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique wood/charcoal the type of woody biomass is less relevant. The estimate is in line with the net annual increments given in Chapter 4 of (IPCC 2006).

4.6.2 Land use modelling Mozambique Initial land use map and modelled land use classes There are several sources for land use maps (JRC 2000, ESA GloBCover project 2005, DNTF et al. 2008); however, there are significant inconsistencies between these maps. These differences are caused by the types of satellites used, the recording periods (years and seasons), the interpretation of the observed reflection spectra and the land use classification. In this study, the Global Land Cover map (2005) was used as the basis for the spatial assessment of current land use and land use developments. The initial land use map with a spatial resolution of approximately 300 m was aggregated to a grid cell size of 2 1 km for practical modelling reasons. Firstly, a higher spatial resolution would result in time-consuming model runs because of the required calculation time; secondly, as most spatial data sets used in this study have a lower spatial resolution, more detailed modelling would give a false impression of accuracy. As many land use classes remain stable over time and some land use classes have similar land use functions, it was feasible to reduce the initial number of land use classes from 26 to 10 without limiting the functionality of the model. The land use classes cropland, mosaic cropland-grassland, mosaic cropland-pasture, pasture and forest were modelled dynamically as they are related to the demand for crops, pasture and wood. Table 4.5 shows the land use classes, the way they are modelled (statically/dynamically), their composition (for mosaic land use classes), and the proportion of the crop and pasture demand they fulfil. In this study, cropland is one homogenous land use class in which no individual crop types are distinguished. There are several reasons: firstly, the ESA land use map does not distinguish between crop types. As there is already considerable uncertainty in the interpretation of satellite images and the translation of reflections into specific land use classes, and because of seasonal changes in land coverage, a further disaggregation is not feasible. Secondly, the spatial distribution of crops in a specific timeframe is not relevant: as crops are generally cultivated in rotations, the spatial distribution of crops changes during and over the years. Thirdly, as all cropland is excluded for biofuel production, a distinction in the spatial distribution of specific crops within cropland is not expected to influence the land availability for biofuel crops. The same applies to the mosaic land use classes. Although the composition of mosaic land use types is spatially heterogeneous, no spatial differentiation has been made in the proportion of cropland and pasture. This is not possible due to the type of classification in GlobCover, the fluctuations over time, and the difficulty in incorporating this into the spatio-temporal modelling.

153

SHADES OF GREEN

Dynamic modellinga

Area (share) b

Proportion cropland c

Proportion pasture c

Share of total crop demand d

Share of total pasture demandd

Table 4.5: Aggregated land use types included in the land use change modelling, the area of these land use types, the proportion of pasture and cropland in these land use types and the proportion of crop and pasture demand that is met by the specific land use types in 2005.

D-dd

0.01

1.00

-

0.13

-

Mosaic cropland - pasture

D-dd

0.00

0.64

0.36

0.87

0.82

Mosaic cropland - grassland

D-dd

0.16

0.64

-

0.00

-

Mixed pasture

D-dd

0.01

-

1.00

-

0.18

Mixed grassland

D-p

0.01

-

-

-

-

Land use type Cropland

Shrub land

D-p

0.20

-

-

-

-

Forest e

D-p

0.59

-

-

-

-

Deforested area e

D-dd

0.00

-

-

-

-

Urban area f

F

0.00

-

-

-

-

Excluded areas g

F

0.02

-

-

-

-

a

This column indicates whether the land use class is dynamically modelled (D) or is included as a fixed area (F). The dynamically modelled land use classes (D) can change over time as the result of a change in demand (demand driven –dd) or as a result of the expansion of other land use classes (passive –p). See section 2.3. b This indicates the proportion of the total land area taken up by the specific land use class. It is based on the aggregation of land use classes of GlobCover in 2005. c The proportion of the area within a land use class, which is dedicated to crop cultivation or pasture. d The proportion of demand (for crops and pasture) fulfilled by specific land use is calculated based on the land use map, the productivity map, the management level, and the total demand for 2005. Over time, a shift towards more dedicated land use is expected (dedicated cropland and dedicated pasture at the cost of mixed land uses). As the Progressive scenario assumes a rapid progress in improved cultivation, the shift towards more commercial farming and therefore dedicated land use types is more profound. The rate of change towards dedicated land uses is related to the weighted average of the annual growth factor of yield of the BAU and the Progressive scenarios. e The land use class ‘forest’ consists of open and closed deciduous and evergreen forest. The current land cover map does not distinguish deforested land. It is included in other land use categories such as shrub land. Forest is a passive dynamic land use class: there is no allocation based on demand for forest, but it could change to other land use types (cropland and pasture) when these categories expand. The category ‘deforested area’ is demand-driven dynamic, as it meets the demand of illegal harvested fuel wood and industrial round wood. f Although the increase in population and the trend towards urbanisation would suggest land use changes towards urban areas, urban areas are not dynamically modelled. The main reason for this is that the limited fraction of the total country area will not significantly influence the land availability for bioenergy crops. g The excluded areas comprise all land use classes that are assumed to be static over time and cannot change to other land use functions. These include inland water bodies, regularly flooded areas (fresh and saltwater) and consolidated bare areas.

154

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique

As the starting point for the LUC modelling, the land use map of 2005 was calibrated with the statistical data of the demands for crops, fodder and grassland of 2005. This means that the projected demands for products from a certain land use class in 2005 should match the production of this land use class in the initial land use map based on the potential yield, given the productivity of the cells’ location and the management level of 2005. Several maps on the productivity of individual crop types or on arable crops in general are available (FAO and IIASA 2000b; MNP 2006; You 2006; FAO - GIS UNIT 2007; 3 DNTF et al. 2008) . In this study, the yield map of FAO and IIASA was selected to estimate the productivity of pastures and cropland. The biomass density map provided by DNTF was used to estimate the productivity of forest. Both maps have a relatively coarse resolution (5 arcmin and 1:1,000,000). The statistics on demand and the land cover maps are outdated and the quality is relatively poor, which complicates calibration. These inconsistencies in data sources were also found in other studies on spatial LUC modelling (Zuidema et al. 1994; Alcamo et al. 1998; Verburg et al. 1999; Schmit et al. 2006). In order to calibrate the land use map with the statistics, this map was combined with the potential yield map of the land use class, in order to assess the total production of each land use class. From there, the proportions could be assessed of the demand for crops, grazing and wood met by a specific land use type (cropland, mosaic cropland-pasture, mosaic cropland-grassland, mixed pasture, forest), and the proportions of cropland and pasture within mixed land use classes. Land allocation The suitability factors applied in this model to allocate land to a certain land use class were determined by expert judgment and literature review. For each land use class, a specific set of suitability factors was selected. Table 4.6 presents these suitability factors, the direction of their relation, the type of correlation and the maximum distance of effect for all land use classes. The order of allocation applied in this study is (1) cropland, (2) mosaic cropland-pasture, (3) mosaic cropland-grassland, (4) pasture and (5) forest (deforestation does not compete with other land use functions). This order is based on the assumed economic importance of these land use classes from the individual farmer’s perspective.

3

There are significant inconsistencies in these maps, not only in relative productivity but even in recognisable patterns of productive/non-productive areas. Inconsistencies are mainly caused by the different environmental criteria included (e.g. soil, slope, precipitation and temperature), the different source data used for the individual parameters and the different interpretations of crop requirements. 155

SHADES OF GREEN

Table 4.6: Characteristics and weights of the suitability factors for land use allocation. Land use type pure cropland and

Suitability factor

Distance of effect Relation

Weight 0.20

no. of neighbours same class

Positive

linked cell

linear

cropland-grassland b

distance to roads

Negative

5 km

inv proportional 0.10

c

distance to water

Negative

10 km

inv proportional 0.10

d

distance to cities

Negative

50 km

inv proportional 0.10

e

yield

Positive

exponential

0.20

f

population density

Positive

proportional

0.20

i

conversion elasticity

0.10

Total

1.00

pasture

cropland-pasture

deforestation

a

a

Direction

a

no. of neighbours same class

Positive

linked cell

linear

b

distance to roads

Negative

5 km

inv proportional 0.05

0.20

c

distance to water

Negative

10 km

inv proportional 0.25

d

distance to cities

Negative

50 km

inv proportional 0.05

e

yield

Positive

exponential

0.10

f

population density

Positive

proportional

0.05

g

cattle density

Positive

proportional

0.20

i

conversion elasticity

0.10

Total

1.00

a

no. of neighbours same class

Positive

linked cell

linear

b

distance to roads

Negative

5 km

inv proportional 0.10

0.20

c

distance to water

Negative

10 km

inv proportional 0.10

d

distance to cities

negative

50 km

inv proportional 0.10

e

yield

positive

exponential

0.15

f

population density

positive

proportional

0.15

g

cattle density

positive

proportional

0.10

i

conversion elasticity

0.10

Total

1.00

b

distance to roads

negative

5 km

Inv proportional 0.25

d

distance to cities

negative

50 km

Inv proportional 0.20

e

yield

negative

exponential

0.05

f

population density

positive

proportional

0.30

h

distance to forest edge

positive

exponential

0.20

Total 1.00 The suitability factor ‘number of neighbours in the same class’ refers to the amount of grid cells in the immediate surrounding that are already part of the same or a related land use class. For cropland, the land use classes cropland-pasture, cropland-grassland and pure cropland are related land use classes. For pasture, both pure pastures and cropland-pasture are related land use classes. The 3x3 Moore neighbourhood (Codd 1968) has been applied here: statistics are calculated for the eight cells surrounding a central cell. For example: for the land use change towards cropland, the cells are counted in the immediate surroundings that are already cropland, cropland-pasture and cropland-grassland. As a result, cropland is more likely to expand where it 156

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique borders on pure or mosaic cropland than at a random location. This neighbourhood function is often used in cellular automata models and is increasingly implemented in LUC models (Verburg et al. 2004). Proximity to same land use or neighbourhood functions is also applied by (e.g. Alcamo et al. 1994; Verburg and Overmars 2009; Lapola et al. 2010). As land use changes over time, the number of neighbouring cells in the same class is assessed for each individual time step (year). The land use map that results from the allocation step t is used for the suitability factor ‘number of neighbouring cells in the same land use class’ for t+1. b The distance to roads is based on the map of road infrastructure provided by ANE (2010). Only primary and secondary roads were selected for this assessment. It is assumed that land close to roads is more susceptible to LUC, as these areas are more accessible for the population. Including this suitability factor ensures that areas close to road infrastructure are more likely to change to man-made land use types than more remote areas. Moreover, in other LUC studies, accessibility was perceived as an important suitability factor, (e.g. Verburg et al. 1999; Lambin et al. 2001; Wassenaar et al. 2007; Lapola et al. 2010), specifically for deforestation, see e.g. Mertens and Lambin (1997; 1999) and Mas et al. (2004). The distance of 5 km is indicative and is based on the findings of Jansen et al. (2008), who found that the majority of land use changes are found within a 10 km buffer from road infrastructure (5 km either side) in the Manica province. c The distance to water resources was calculated based on the map provided by the National Directorate of Land and Forests (DNTF 2008e), which mapped the main water resources (rivers and lakes). It is assumed that pasture areas and cropland are more likely to expand in areas with access to water than in areas without water resources. The proximity to the nearest river is also used as suitability factor in the study of Verburg et al. (1999), and it was indicated as an important factor during expert consultation in Mozambique. The distance of effect is set at 10 km, based on the distance that is commonly covered on foot by women to fetch water in Rural Africa (UNDP 2004). d The distance to cities expresses the pressure from urban areas on the surrounding land. In addition to the local population density, it accounts for the additional pressure caused by the proximity of markets. Only the province capitals (based on the map provided by ANE) were included in this assessment. Mertens et al. (1997) also included cities that exceeded a certain size. In several LUC models, proximity to settlements or to markets was included as suitability factor (e.g. Wassenaar et al. 2007; Walsh et al. 2008; Lapola et al. 2010). Here, the distance of effect has been set at 50 km based on the dominance of local markets over regional and national markets for commodities. e The yield maps provide information on the relative productivity of cropland, pasture or forest. The incorporation of this suitability factor ensures that the expansion of cropland or pasture is more likely to expand in areas with high potential yield levels, compared to areas with marginal potential production. The relative yield map is derived from FAO and IIASA (2000b), and DNTF (2008) has provided the biomass density map that provides information on the potential amount of wood to be extracted from forest. The productivity or potential yield levels are frequently used as suitability factor in LUC models, e.g. Alcamo et al. (1994), Veldkamp et al. (1996), Verburg et al. (1999), Lapola et al. (2010). f The general pressure on land related to population is expressed in the suitability factor ‘population density’. The population density map was retrieved from FAO (2000). The suitability factor ‘population density’ has been included because deforestation and the expansion of cropland and pasture are more likely to occur in areas with a high population density. Although progressive urbanisation is assumed towards 2030, no estimations were made about shifts in spatial distribution of the population due to a lack of spatial data. The suitability factor population density is frequently used in LUC models, see e.g. Verburg et al. (1999) and Wassenaar et al. (2007). g ‘Cattle density’ is an important factor for the expansion of pasture land. As both goats and cattle have grazing requirements, the spatial distribution of both species influence the expansion of pastures. In order to develop a general map of the spatial distribution of grazing animals, the cattle and goat density maps (FAO and IIASA 2000b) were combined in a weighted summation based on the ratio of feed requirements of cattle and goats. Although the ratio of feed requirements of goats and cattle change over time due to different rates of efficiency improvements, the ratio for the weighted summation has been kept constant because a changing 157

SHADES OF GREEN

h

i

ratio over time would require a different summated input map for every modelling year, which would make the modelling more complex. Including the suitability factor based on cattle density ensures that expansion of pasture areas is more likely to occur in areas with a high density of grazing animals. ‘Distance to forest edge’ is an important suitability factor for the spatial allocation of deforestation as it is assumed that it is more likely that deforestation occurs at the border area between forest and non-forest than in the middle of forest. This suitability factor is frequently used in spatial modelling of deforestation, see e.g. Mertens and Lambin (1997; 1999) and Mas et al. (2004). The suitability factor ‘conversion elasticity’ consists of a matrix which depicts the relative probabilities of the conversions of one land use type to another. Higher values indicate a more probable conversion. For example: in the matrix, it is depicted that mixed cropland-grassland is relatively easily converted into pure cropland but that forest is more difficult to convert. Conversion elasticity is also used in Verburg and Overmars (2009).

Table 4.7 summarises the land use types, land covers, and physical constraints that are excluded for LUC. In addition to the areas excluded for LUC towards agricultural land uses, 4 additional areas have been excluded for the production of bioenergy crops . Table 4.7: Categories of land use types, land cover types and physical constraints excluded from land use changes towards cropland and pasture and excluded from the land available for bioenergy crops. Excluded for Sources of spatial data c bioenergy crops Land use Urban areas √ √ (ESA 2005) Community area √ (DNTF 2008a) Infrastructure √ √ (ANE 2010) Concessions √ (DNTF 2008b) Land use rights √ (DNTF 2008c) Protected areas √ √ (DNTF 2008d) (ESA 2005) d Cropland √ Cropland-pasture √ (ESA 2005) d Cropland-grassland √ (ESA 2005) d Pasture √ (ESA 2005) d Land cover Forest a (√) √ (ESA 2005) d b Deforested areas √ (ESA 2005) d Physical Static, unsuitable areas √ √ (ESA 2005) d Steep slopes (>16%) √ √ (NASA and NGA 2000) constraints a In the BAU scenario it is assumed that cropland and pasture can expand in forest areas (in line with current practices). In the Progressive scenario it is assumed that enhanced policies and increased enforcements can prevent deforestation. Therefore, forest areas are excluded for land use change in the Progressive scenario b Deforested areas are excluded from the land available for bioenergy crops in both the BAU and the Progressive scenario as it should be prevented that the introduction of bioenergy crops becomes an additional driver for deforestation. Category

Excluded land categories

Excluded for agriculture

4

The spatial data used to exclude areas such as community areas or protected areas were derived from Mozambican institutes and verified at a local level (district level). However, the scale of these maps is relatively coarse (1:1,000,000). A new zoning assessment has been started in 2011 to develop spatial data on a scale of 1:250,000. This assessment is expected to be finalised in 2015.

158

4. Spatiotemporal land use modeling to assess land availability for energy crops illustrated for Mozambique c

The spatial data used to exclude areas such as community areas or protected areas were derived from Mozambican institutes and verified at a local level (district level). However, the scale of these maps is relatively coarse (1:1,000,000). A new zoning assessment has been started in 2011 to develop spatial data on a scale of 1:250,000. This assessment is expected to be finalised in 2015. d Output of t is input for t+1

Stochastic model inputs A number of input data and model parameters are uncertain due to a lack of and discrepancies between data sources of LUC drivers in Mozambique. In order to assess the effect of uncertainties (in for example time series of demand and productivity, spatial input maps and the characteristics of suitability factors in the allocation procedures), a Monte Carlo analysis was performed in which some of the input data and parameters were modelled stochastically Table 4.8. In addition, the type of error model and the standard deviation of the error applied are shown. The uncertainty range in demand was introduced to deal with the uncertainties in the projection of population growth and dietary intake per capita. The stochastic input is a uniform distribution between two time series representing the lower and higher estimations of the development of demand. The lower estimations represent the lower projections of population growth (32 million people in 2030; (UNDP 2008) in combination with an unchanged dietary intake per capita compared to the 2005 levels (2050 kcal/cap/day). The upper estimation represents the high population growth projections (36 million people; (UNDP 2008) and a dietary intake in 2030 equal to the average of developing counties in 2030 (2980 kcal/cap/day; (FAO 2003b). The time series of the maximum attainable yield level are scenario-specific and relate to technology adoption. As it is uncertain how agricultural productivity will develop over time, the development of the maximum yield in the two scenarios has been modelled stochastically as a uniform distribution between two time series representing a lower and a higher estimation of the development in productivity for the two scenarios. The lower estimation represents the productivity increase equal to 0.5 times the annual growth rate of the normal productivity increase of the scenario. The upper bandwidth is assumed to be equal to the lower bandwidth.. As yield maps of several sources are contradictory, a large relative standard deviation has been applied for the stochastic input. The dynamics in the spatial patterns in population and livestock densities have not been modelled over time. In order to take into account the uncertainties in the current population and livestock density and the potential changes in spatial distribution, the population and livestock density maps have been modelled stochastically. The characteristics of the suitability factors are uncertain (e.g. spatial data, the relation between suitability factor and effect, the distance of effect in the ‘distance to’ and ‘neighbouring’ function, and the 159

SHADES OF GREEN

weights of the suitability factors); in addition, they may be spatially heterogeneous. To assess the effects of these uncertainties on the result, the characteristics have been modelled stochastically. The error models and the standard deviation were selected based on expert judgement and were used to provide a proof of principle. However, they do not necessarily represent the actual uncertainty range of the parameters. Table 4.8: Stochastic input variables. Type of input

Stochastic input

Error model

Standard deviation 0.1 0.1 0.1 1000 m 10 m 1000 m, which results in a

Driving forces

Demand Uniform Maximum attainable yield Uniform Yield map Relative Suitability factors Map Population density Relative Map Cattle density Relative Neighbourhood functiona Normal Distance to functionb Uniform Elevation mapsc Normal a The neighbourhood function is based on a 3 km window with a normal error of normal distribution between 2 and 4 km window length. b The distance to function is applied for all three ‘distance to’ functions (road, water and cities). The selected upper limit is 200% of the deterministic length (Table 6). The lower limit is one cell length (1000 m) c Based on the Digital Elevation Map (DEM), slopes over 16% have been excluded for agricultural land use and deforestation. By modelling the DEM stochastically, the area excluded because of steep slopes is a stochastic input. A normal distribution and a 10 m standard deviation is based on the findings of Rodriguez et al. (2005).

160

SHADES OF GREEN

162

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

5

Spatiotemporal costsupply curves for bioenergy production in Mozambique F. van der Hilst and A.P.C. Faaij Accepted for Publication in BioFPR Volume 6, Issue 4 doi: 10.10002/bbb.1332

163

SHADES OF GREEN

ABSTRACT The objective of this study is to assess how bioenergy cost and supply potential in Mozambique develop over time in a spatially explicit way. The land availability for energy crops was explored making use of a business as usual and progressive scenario on the development of other land use functions. The assessment of the cost and supply potential is based on the developments in land availability, the suitability of the land that is and becomes available, the disaggregated cost break down of energy crop production, the transportation distance of feedstock to conversion plant, the cost of conversion, the transportation distance from plant to harbour and the cost of international shipping. The supply chains of eucalyptus (torrefied) pellets and sugarcane ethanol are used as a case study. The results show a large spatial variation in supply chain costs which is the result of spatial variation in feedstock production costs, primary transport costs and secondary transport costs. Most promising areas for eucalyptus and sugarcane production are scattered in the central south, the central, and the north eastern part of Mozambique where agro-ecological conditions are relatively favourable, where sufficient feedstock can grow to meet plant input requirements and where infrastructure is available. In 2030 in the progressive scenario, the total potential for eucalyptus pellet production amounts 3200 PJ of which 2500 PJ could be exported to Europe below a market price level of 8 €/GJ and for sugarcane ethanol the total potential amounts 850 PJ of which 500 PJ could be exported below a price level of 30 €/GJ. The location of production is the key factor for cost effective production. This study demonstrates an approach which enables the assessment of the development of bioenergy potentials and costs over time in a spatially explicit way. As environmental and socio-economic impacts of bioenergy supply chains are highly related to the biophysical and socio-economic context of the location of production, a spatially explicit assessment of bioenergy production potential is a suited approach for the assessment of the sustainability of bioenergy chains.

164

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

5.1 Introduction Studies on global bioenergy potential have indicated large techno-economic production potentials in the Southern African Development Community (SADC) regions of Sub Saharan Africa (Hoogwijk et al. 2005; Smeets et al. 2007; Hoogwijk et al. 2009). Mozambique is considered a promising country for biomass production within this region because of the availability of land (Batidzirai et al. 2006; Namburete 2006); the favourable environmental conditions for agricultural production (INE 2003; Batidzirai et al. 2006), and the relatively low current agricultural productivity which offers potential for improvement (World Bank 2006). The main incentives for the government of Mozambique to focus on the development of a bioenergy industry are to decrease the dependence of oil imports (15% of the total national import expenditures (Cuvilas et al. 2010)) and to enhance energy security and socio-economic and sustainable development especially in rural areas (Conselho de Ministros da república de Moçambique 2009). In recent years, an increasing number of studies has been published on bioenergy potentials on a global, European, national and regional level (e.g. Hoogwijk et al. 2005; Smeets et al. 2007; van Dam et al. 2009a; de Wit and Faaij 2010; Dornburg et al. 2010; Haberl et al. 2010; Krasuska et al. 2010). Some studies have assessed the bioenergy production potential in Mozambique (Batidzirai et al. 2006; Econergy 2008; Watson 2011). However, most of these studies have assessed bioenergy potentials on a spatially aggregated level. The disadvantage of these spatially aggregated studies is that they provide limited information on the location of the land available for energy crop production and how this evolves over time. As potential yield levels and production costs of energy crop production are strongly related to the physical and socio-economic conditions of a location (van Dam et al. 2009a; 2009b; Van der Hilst et al. 2010; 2012a), developments in actual potentials and cost can only be assessed spatially explicitly. Van der Hilst et al. (2011) assessed the development of the land availability for bioenergy crops in Mozambique in the timeframe 2005-2030 on a detailed spatial level (cell size 1 2 km ), while taking into account the developments in land use requirements for agriculture and deforestation and accounting for the uncertainties in key determinant factors of land use change. This is an important first step to assess the developments in actual supply potentials and costs of bioenergy chains. The objective of this study is to assess the development cost supply potential of bioenergy crops in Mozambique over time in a spatial explicit manner, while taking into account the developments in land availability, energy crop yield, conversion technology and the

165

SHADES OF GREEN

logistical chain. Two divergent bioenergy supply chains are assessed in this study: (torrefied) pellets from eucalyptus and sugarcane ethanol. These bioenergy chains are selected because of the potentially high crop yields, the promising economic performance, the experiences with these crops in Mozambique, the different crop requirements, their divergent properties and (multiple) end products (Batidzirai et al. 2006; Watson et al. 2008).

5.2 Methodology and data input The cost developments of eucalyptus pellets and sugarcane ethanol are calculated 2 spatially and temporally explicitly at a cell size of 1 km and at a yearly interval for the time frame 2010-2030. The costs included are feedstock production cost, primary transport (from field to plant), conversion (to ethanol) or pre-treatment (to pellets), secondary transport (from plant to harbour), storage and international shipping. All costs are 5 calculated in €2010 and indexed for inflation.

5.2.1 Land availability The developments in land availability for bioenergy crops in Mozambique in the timeframe 2005-2030 have been assesses in the study of van der Hilst et al. (2011). As it is of key interest that competition for land and related effects of indirect land use change (iLUC) are avoided, the land availability for energy crop production was modelled while taking into account the land required for other land use functions such as nature conservation and food production. The demand for domestically produced food and feed is related to developments in population size, Gross Domestic Product (GDP), food intake per capita and self sufficiency ratio (SSR, i.e. the extent to which domestic supply meets domestic demand (FAO 2003b)). The efficiency of the agricultural sector is a key factor for the land required to meet the total demand for food, animal products and materials. A scenario approach was used to explore potential long term developments in the productivity of the agricultural sector. The Business as Usual (BAU) scenario projects a future in which historical trends in yield levels and livestock productivity are continued resulting in a low agricultural productivity. The progressive scenario represent a discontinuation of historical trends: it assumes the implementation of improved agricultural management resulting in a high agricultural productivity. The land use changes 2 in the timeframe 2005-2030 were modelled for each year on a 1km grid cell size level by allocating land to a land use class based on the suitability for the specific land use classes. The suitability of land was defined by a spatial weighted summation of a specific set of 5

A discount rate of 12% is applied and exchange rates are set on 2010 averages: 1€ = 1.32 US$ and 1€ = 40.73 MZN. 166

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

suitability factors (i.e. the vicinity of the same land use class; the productivity; the distance to road, water and main cities; population and cattle density; conversion elasticity; and the distance to forest edge). Areas that are not suitable (such as steep slopes) or not allowed (such as conservation areas) to be converted to agricultural land, were excluded. Based on the allocation of land use classes and the maps of excluded areas for bioenergy production (such as community land), the land availability for bioenergy crops was determined for each year. Table 5.1 provides a summary of the amount of available land in for the two scenarios in the timeframe 2010-2030. The land availability for energy crops decreases over time in the BAU scenario, due to an increased demand for food, feed and fibre (due to an increase in population and consumption per capita) on the one hand and a very minor increase in productivity of crop and livestock production on the other. In the progressive scenario, the land availability for energy crops increase as the land requirements for agricultural production are reduced as a result of an intensification of the agricultural sector. (DNTF et al. 2008) The technical characteristics of the PC Raster Land Use Change model (PLUC) developed for the land use allocation are found in Verstegen et al.(2011). The methods and the data inputs for the modelling and the resulting maps of the development in land availability for bioenergy crops in the business as usual and the progressive scenario for the timeframe 2005-2030 are available from the study of van der Hilst et al. (2011). Table 5.1: Land availability for bioenergy crops in Mozambique in the timeframe 2010-2030 for the 6 business as usual and the progressive scenario. Land availability Business as usual scenario Progressive scenario

Unit Mha Mha

2010 8.89 10.45

2020 8.37 13.61

2030 7.72 16.41

5.2.2 Feedstock production costs Feedstock production costs are assessed by calculating the net present value (NPV) of all costs items and the biomass yield during the lifetime of the biomass production plantation. This method has frequently been used for the calculation of the costs of (perennial) biomass feedstock production (e.g. van den Broek et al. 2000b; van Dam et al. 6 The amount of land is indicated as available in 2010 is 28% higher than indicated by the zoning assessment commissioned by the Mozambican Government and performed by DNTF, IIAM and CenaCarta (2008). The differences are caused by the difference in recording period of the satellite images, in interpretation of the satellite images, in the mapping method (vector vs grid), in the resolution, and in the areas excluded. The scale of the zoning assessment was relatively coarse (1:1,000,000). A new zoning assessment has been started in 2011 to develop spatial data on a scale of 1:250,000. This assessment is expected to be finalised in 2015.

167

SHADES OF GREEN

2009a; Van der Hilst et al. 2010). Some of cost items are costs ‘per hectare’ such as cost for land, land preparation and pesticide application. Other cost items are related to the production volume such as the costs for fertilizers (as application levels are linked to 3 nutrient removals) and harvest costs (per m or ton harvested). Equation 5.1 provides the method used to calculate the discounted cost per oven dried tonne (odt) for eucalyptus or ton cane (TC) for sugarcane. N

∑(I

⋅C

M

) + ∑( J

⋅C

⋅Y

)

 Yy /∑  (1 + a )y 

ny ny my my y y =x Y =x n 1= m 1 = cr y Y 1 =x 1

C =∑

(1 + a )

    Equation 5.1

Ccr I Cny J Cmy Y a y

Discounted costs feedstock production occurrence cost item per ha n in year y cost of cost item n in year y occurrence of cost item per odt m in year y cost of cost item m per odt yield in year y discount rate annuity period (lifetime plantation)

€/odt # €/ha # €/odt odt/ha % y

The data required for the cost calculations are derived from extensive literature review and field visits to several plantations in Mozambique. Due to variations in agro-ecological conditions, the yields and related production costs of energy crops are spatially highly heterogeneous. In order to calculate the spatially explicit feedstock production cost, the map of land availability of year y is combined with the crop suitability map and the maximum attainable yield given the level of management in year y.

Yay = Aay ⋅ Sa ⋅ My Equation 5.2 Yay Aay Sa My

Yield of energy crop at location a in year y Land availability of lactation a in year y Suitability of land at location a Maximum yield given management level in year y

ton/ha I /0 % ton/ha

The bioenergy production costs are expected to change over time due to technological learning and adoption of more advanced management practices. In studies on the cost development of sugarcane in Brazil (van den Wall Bake et al. 2009), corn in the US (Hettinga et al. 2009), and rapeseed in Germany (Berghout 2008), the experience curve approach was applied in which the development in costs are related to the cumulative

168

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

production (Boston Consultancy Group 1972). De Wit et al. (2011b) demonstrated the possibilities and limitations of using the experience curve approach for the assessment of future developments in the feedstock production costs of woody biomass. These studies showed that an increase in productivity is the most important driver for decreasing production costs for feedstock, contributing between 65 and 85% to the total cost decline. The additional 15-35% cost reductions in production costs are the result of increasing economic efficiency in production operations e.g. increased mechanisation and tailored pest control. However, these developments could be counteracted by future price increases of inputs such as diesel, fertilizers and chemicals (van den Wall Bake et al. 2009). Because these mechanisms are quite uncertain and because increased productivity shows the results of improving management, yield development is selected as the most suitable parameter for the assessment of cost developments over time (de Wit et al. 2010). It is assumed that dedicated biomass plantations in Mozambique are established by (foreign) investors which apply state of the art management techniques based on global experience with dedicated biomass plantations. This is in line with current investment projects in Mozambique (República de Moçambique 2010). The cost of feedstock production in Mozambique are variable due to variations in agro-ecological conditions and due to differences in economies of scale, plantation designs, management practices, local variations in input cost, and geographical location. The cost estimates for this study are specific for a state of the art plantation and are based on extensive literature review and field visits to plantations in Mozambique. For eucalyptus a total lifetime of 21 years with a coppice cycle of 7 years is assumed. This is equal to the coppice period currently used for the paper and pulp industry in Mozambique (field work) and in Brazil (Walter et al. 2006). For sugarcane a total lifetime of 24 years is assumed, with 4 cycles of plant cane and 5 ratoons. Table 5.2 present the cost of feedstock cultivation of Eucalyptus and sugarcane in Mozambique. The development in yield is selected as the most suitable parameter for the assessment of cost developments over time. The maximum attainable yield is expected to increase due to technological learning and is based on yield levels currently achieved in Mozambique as well as on historical and expected growth rates. The average yield levels are derived from the land availability maps for the Progressive and the Business as usual scenario up to 2030 produced in the study of van der Hilst et al (2011) combined with the suitability maps provided by FAO and IIASA (2000a; FAO - GIS UNIT 2007). For each grid cell, these maps provide the percentage of maximum attainable yield that can be achieved. The average yield levels are much lower as the average suitability of land available in 2010 is 35% of the maximum attainable yield for eucalyptus and only 17% for sugar cane. This low 169

SHADES OF GREEN

average suitability of sugar cane is because only a minor part of the available land is suitable for sugar cane cultivation. The majority of the available land is not, or only marginally suitable for sugar cane cultivation. See Table 5.3 for the development of maximum and average yield levels up to 2030. Table 5.2: Cultivation cost items of eucalyptus and sugarcane in Mozambique. Cultivation

Unit

Eucalyptus

Land lease a

€/ha/y

0.10

Land clearing b

€/ha

946

€/ha

Reference s a

Sugarcane 0.74

Reference s a

c

946

c

-

-

401

c

€/ha

256

c

256

c

Irrigation instalment b

€/ha

-

-

1891

c

Irrigation

€/ha/y

-

-

436

c

Planting d

€/ha

24

c

31

c

€/ha

188

c

250

e

€/odt - €/TC

5.94

f, g

1.98

c, f

45.13

c

Land forming b Field bed preparation

Seeds

b

b

Fertilizers Chemicals

€/ha/y

19.00

h, i

Agro chemical application

€/ha/y

11

i

138-311l

c

9.77

j, k

2.64

c

Harvesting + Loading

€/odt - €/TC

a

The cost of land lease consist of cost for initial authorisation of 21.10 €/ha and a annual fee 0.10-0.74 €/ha depending on the purpose of the land. Based on (CPI 2009). b it is assumed that land clearing, land forming, and irrigation instalment is required once (at the start of) the lifetime of the plantation. Field bed preparation and planting is required once for eucalyptus and once every 6 years for sugar cane. c (Açucareira de Xinavane SA 2010) d Seedling costs based on 3x3 planting density, 15% mortality, 0.15 €/seedling. Based on field visits three plantations in Mozambique e (Salassi and Deliberto 2010) f (Chemonics and IFCD 2007) g (Laclau et al. 2003) h (van den Broek et al. 2000b) i (Econergy 2008) j (Hamelinck et al. 2005b) k (Savcor 2006) l These are the cost for labour, machinery and fuel for plant maintenance. In the year of planting it amounts 138 €/ha, and in the 2nd - 6th year the costs are 311 €/ha.

170

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

Table 5.3: Maximum and average yield levels of eucalyptus and sugarcane in Mozambique up to 2030 given the land availability in the Business As Usual and the Progressive scenario. Crop

Yield level

Eucalyptus a

Maximum Average available land BAU

Sugarcane b

a

b

2010

2020

2030

odt/ha/y

22.6

26.3

30.5

Growth rate % p.a. 1.50

odt/ha/y

7.9

9.1

10.4

1.39

Average available land PROG

odt/ha/y

7.9

9.6

11.6

1.95

Maximum

TC/ha/y

140

149

158

0.60

Average available land BAU

TC/ha/y

23.8

25.3

27.1

0.66

Average available land PROG

TC/ha/y

24.0

27.5

31.2

1.33

There is a large range in reported and expected yield levels of eucalyptus. Estimates of the mean annual increment (MAI) of eucalyptus in Mozambique and Sub Saharan Africa vary between 4.5 to 35.0 odt/ha (Ugalde et al. 2001; IPCC 2003; Laclau et al. 2003; Batidzirai et al. 2006; Savcor 2006; van Eijck et al. 2012) and field visits to 3 plantations in Mozambique). The annual growth in MAI is estimated on 1.5% per annum and is based on a feasibility study of an eucalyptus plantation in Mozambique (Savcor 2006), and projected yield increases of perennial cash crops in Mozambique (Van der Hilst et al. 2011). The projected maximum attainable yield in 2030 is still well below the estimated theoretical maximum yield for Mozambique (Batidzirai et al. 2006). The average yield level of the available land in the BAU en PROG scenario is based on van der Hilst et al. (2011). Sugarcane yield levels decrease during the ratooning period. Van der Wall Bake et al. (2009) reported a yield decline of 15% after the first harvest and 6-8% in the subsequent years. In this study, an average yield decline of 4.5% p.a.is assumed based on the figures reported by one of the Mozambican sugar mills (Açucareira de Xinavane SA 2010; De Vries et al. 2011a) and in line with by the report of Econergy (2008). A yield level of 140 TC/ha evolves as follows: 140; 134; 139; 123; 116; 111 TC/ha. The average suitability of the available land in the BAU and PROG scenario is based on (Van der Hilst et al. 2011).

5.2.3 Conversion costs The conversion costs comprise investment costs, operation and maintenance (O&M) costs, and energy input costs. For the eucalyptus supply chain, only the pre-treatment step is included as pellets are assumed to be the end product that are exported for cofiring for electricity production elsewhere. It is assumed that the costs of pre-treatment/ conversion are not location specific and are therefore not calculated spatially explicitly. It is assumed that in the short term wood pellets are produced in the eucalyptus supply chain and that there will be a gradual shift towards torrefied pellets (TOP) when this technology becomes more mature. Pelletising is a commercially applied technology which consists of four main steps: drying, grinding, densification and cooling. The overall thermal efficiency is 92.2% (excluding utility fuel) (Uslu et al. 2008). Torrefaction technology is not commercially available yet. In the torrefaction process assumed in this study, the combustion of torrefaction gas is expected to cover the energy demand of the dryer resulting in system with an efficiency of 95% (Uslu et al. 2008). The characteristics of the palletising and TOP production plant are derived from Uslu et al. (2008) who assumed a 171

SHADES OF GREEN

full commercial plant and no economies of scale beyond a scale of 40 MWth(in). Ethanol production from sugarcane is a relatively well established technology. The production process consists of washing, milling, extraction, purification, fermentation and distillation. The efficiency of the fermentation process ranges from 80-90% (van den Wall Bake et al. 2009). The bagasse resulting from the milling is used to feed the boilers to produce steam and electricity for the production process. Surplus electricity could be produced (depending on process and boiler efficiency) to be fed into the grid. Table 5.4 provides the pre-treatment and conversion plant characteristics for the short (2010) and the longer term (2030). Table 5.4: Conversion plant characteristics for wood/torrefied pellets and ethanol production for short (2010) and long (2030) term. Plant characteristics

Capacity

Unit Short term Wood pellets MWth(in)/y

Pellets a

40

Long term TOP

Unit Short term

Ethanol b Long term

40

MWth(in)/y

840

1000

Load factor

h/y

8000

8000

h/y

4380

4380

Efficiency

GJlhv (out) /GJlhv (in)

0.92

0.95

l/TC

85

85

M€

114

107

€/kWth

17

17

€/GJfeed

0.74

0.74

Capital investment

M€

6.2

7.8

Specific investment

M€/MWth

0.15

0.2

O&M

%

5

5

Electricity

Use kWh/odtin

184

146

Prod KWh/TC

90

126

Depreciation period

y

15

15

y

15

15

c

End product characteristics

d

Energy density

GJlhv/odt

18.4

21.5

Moisture content

%

5

10

Mass density

Ton/m3

0.6

Input characteristics Moisture content harvest Moisture content dried in field

% %

30

Energy content

GJ/ton

18.4

0.8

Wood

MJlhv/l Ton/m3

e

55

a

21.1 0.789 Sugarcane f

%

73

GJ/TC

5.3

Data of wood pellet and torrefied pellet production is based on the study of Uslu et al. (2008). A scale of 40 MWth is assumed as larger scales will not lead to economies of scale according to Uslu et al. (2008). b Data of short term sugarcane ethanol production is based on the cost factors of provided for Brazil (van den Wall Bake et al. 2009). These costs are lower than the cost reported for the currently build ethanol plant in Mozambique (of 220 M€ for a 840 MWth plant)(Principle Energy 2009; Reuters 2009). However it is not clear what is included in these investment costs and what margins are included. For that reason, the cost figures of long term assessments of bioethanol plants in Brazil are used. The figures for ethanol production on the long 172

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

term are based on (BNDES and CGEE 2008; Zuurbier and van der Vooren 2008; van den Wall Bake et al. 2009; IPCC 2011; Seabra and Macedo 2011). c Based on (Uslu et al. 2008; de Wit and Faaij 2010) d Based on (Murphy and McCarthy 2005) e Based on (Batidzirai et al. 2006; Wicke et al. 2009; de Wit and Faaij 2010) f Based on (Zuurbier and van der Vooren 2008; IPCC 2011)

Due to the shift from wood pellet to TOP production, the process efficiency increases from 92.2 to 95%. Although the conversion costs of TOPs are slightly higher compared to wood pellets, the characteristics of TOPs are more advantageous for transport and co-firing. A shift from conventional energy inputs (electricity from the grid) to a self sufficient pre treatment plant could be expected (power and heat production by means of a wood fired CHP plant). Although this has benefits from a environmental point of view, it will not necessarily result in cost reductions. Ethanol production costs (excluding feedstock cost) declined by a factor of three between 1975—2005 mainly because of increasing scales and load factors of ethanol plants (van den Wall Bake et al. 2009; Faaij and Junginger 2010). It is expected that costs continue to decrease in line with the learning curve approach of van de Wall Bake et al. (2009) and cost estimates of IPCC (2011). It is assumed that the milling and boiler efficiency will increase resulting in higher electricity revenues in line with the projections of ethanol process developments (Macedo and Seabra 2008; Seabra and Macedo 2011).

5.2.4 Transportation costs Biomass logistics contribute significantly to the total cost per GJ bioenergy produced and delivered (Dornburg and Faaij 2001; Hamelinck et al. 2005b). Key factors of determining the cost of primary transport are the scale of conversion plant and the biomass availability in an area. The cost of transportation of end products from the conversion plant to the harbour depends on the spatial distribution of biomass production and the availability and the quality of road infrastructure. For these reasons, the costs of primary and secondary transport are spatially highly heterogeneous. There is a trade off between minimizing the transport distances of the low density raw feedstock and minimizing the conversion cost due to economies of scale. The optimization of the feedstock transportation distance depends on the scale factor (r) of the technology, the required supply radius (due to distribution and productivity of available land), and the availability and quality of infrastructure. The average required transportation distance from any location a to a conversion plant of scale I in year y can be

173

SHADES OF GREEN 7

calculated according Equation 1 . The required transportation distance changes over time, as the scale of the conversion plant, the crop management factor and the land availability changes over time. −0.5 2 TC ay = C y ⋅ Iy 0.5 ⋅ ( Aa ⋅ Sa ⋅ My ⋅ Day ) 3

Equation 5.3 TCay Cy I Aay Sa My Day

Transport costs from field to plant at location a year Cost of off-road transport Input conversion plant in year y Land availability of location a in year y Suitability of land at location a Maximum yield given management level in year y Distribution of available land at location a in year y

y

€/ton €/tonkm ton/yr 1 or 0 % ton/ha %

The last part of Equation 5.3 (Aa·Sa·My·Day) can be calculated making use of the outcomes of the calculations according to Equation 5.3 (calculation of yield at location a in year y) and the focal statistic tool in ArcGIS. This enables the calculation of the biomass density in 2 ton/km in a certain radius from each grid cell, the required biomass gathering area, and transport distance (see Figure 5.1, left side).

Figure 5.1: Left side: the required biomass gathering area given the input requirements of the conversion plant, the distribution of available land, the productivity of the land and the 7

The exact location of the plant is not decided (it depends on several factors and the choice of location could be a result of this study). A grid cell with feedstock production could supply the feedstock to an abundant amount of potential plants located in any grid cell in the vicinity. The average distance from the field to any of this potential plants is assumed to be the same as the feedstock gathering radius that would be required to meet the plant input demand if the plant was located in that particular gridcell. Given the input requirements of the plant (ton feedstock per year) and the feedstock density (given the availability and the productivity of the land in the surroundings), the average required radius of feedstock gathering can be calculated for that particular grid cell. The average transport distance is 2/3 of the required gathering radius.

174

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

management factor. The average transport distance is 2/3 of the radius of the required gathering area. Right side: the least cost distance from any grid cell to a harbour given the cost surface determined by the availability and the quality of road infrastructure.

To calculate the transportation cost from the conversion plant to the harbour, the logistics are assessed spatially explicitly taking into account the availability and the condition of current infrastructure. In this study, only road infrastructure is included as currently the capacity and reliability of railroad is low (Meeuws 2004) and waterways are not used for large scale transport (yet). Making use of GIS, the least accumulative cost distance from 2 each grid cell (1km ) which is indicated as available for energy crop production to the nearest harbour is calculated over a cost surface (see Figure 5.1, right side). The cost surface consist of a value for each grid cell which represents the cost per unit distance for transport trough the grid cell, based on the spatial information of the availability and the condition of road infrastructure and the related cost per tonkm. The least cumulative costs are calculated for each cell using GIS, summating the costs of each cell that is 8 crossed en-route from each cell to the harbour (corrected for diagonally crossing). Although large scale rehabilitation of destroyed infrastructure is achieved, the condition of the road network is generally poor in Mozambique (Meeuws 2004; USAID 2008). Therefore, road transport is often time consuming and expensive. Costs of transport vary widely due to spatial variation in road condition and accessibility. Reported costs generally range between 0,019 and 0,150 €2010 per tonkm (Meeuws 2004; GDS 2005; Tostao and Brorsen 2005; Savcor 2006; Econergy 2008; USAID 2008) for both liquids and dry bulk. It is assumed that for local feedstock transport from the field to the plant, the higher end of the cost range is most applicable as infrastructure is often lacking in the more remote areas. In order to calculate the transport distance from the field to the plant, the feedstock input requirements in ton/yr are calculated from the plant characteristics depicted in Table 5.4. In addition, the biomass density within a radius of 10 km for eucalyptus and 20 km for sugarcane is calculated for each grid cell for each year. The spatial information of the road network is derived from ANE (National Road Administarion of Mozambique, 2010). Based on the categorization of the road network (primary, secondary, tertiary and vicinal), the type of roads (paved, gravel, earth) and the condition of the road (good, fair, poor, very poor) in combination with recent road maps and own 8

As the exact location of the pre-treatment/conversion plant is not decided, the actual distance between plant and harbour can’t be calculated. However, the grid cell with feedstock supplies to a pre-treatment/conversion plant which is located in a radius around the location of the field. This plant can be located north, south, west or east from the location of the field. Therefore the distance between the plant and the harbour could be both longer or shorter compared to the distance from the field to the harbour. Taking all possible locations of the plant into account, the average distance from plant to harbour will approximately be the same as the distance from the field to the harbour. 175

SHADES OF GREEN

experience, cost per tonkm are assigned to the several types of roads. In Table 5.5, the road classification and the cost per tonkm allocated to them are presented. Table 5.5: Transport cost in €/tonkm for several road types and conditions. Type of road

Condition

paved

gravel

earth

Primary roads

good

0.02

0.04

0.06

fair

0.04

0.06

0.08

poor

0.06

0.10

0.10

very poor

0.10

0.12

0.12

good

0.04

0.06

0.08

fair

0.06

0.08

0.10

poor

0.10

0.10

0.12

Secondary roads

0.12

0.12

0.12

Tertiary roads

very poor

0.06

0.08

0.12

Other roads

0.10

0.12

0.12

Figure 5.2: Road infrastructure Mozambique, derived from ANE (2010).

176

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

Considering the density of the wood /torrefied pellets and ethanol, the mass capacity of 9 the truck is the restrictive factor for transported volume. It is assumed that the ton km price is equal for dry bulk and liquids. Storage of pellets or ethanol is required when there is a difference in the window and volume of production and the frequency and capacity of the downstream logistical chain (Hamelinck et al. 2003). The location, size and costs of storage facilities depend on the biomass delivering schedule and the design of the logistical chain. In this study, a temporary storage in the harbour before international shipment is assumed. Freight rates for international shipping of biomass depend on several factors and can fluctuate considerably over a short period of time (Bradley et al. 2009; Obernberger and Thek 2010; Sikkema et al. 2011b). For the purpose of this study, the cost of shipping are calculated (which are significantly different from short time charter prices) based on the investment costs, O&M costs, lifetime, heavy fuel use, fixed trip costs, and shipping and (un-) loading time. To deal with the uncertainties in shipping costs, variations in costs are incorporated in the sensitivity analysis. Table 5.6: Transport and storage characteristics of wood pellet and ethanol for road transport and international shipping. Transport characteristics

Unit

Pellet

Ethanol

Reference

Truck capacity

ton/truck

40

25

b

m3/truck

130

33

b

3

0.55

0.55

b

2

1.6

b, c

Load-unload costs truck

€/m

Load cost port

€/ ton - €/m3

Shipping costs a

€/ton - €/m3

15.19

Ship capacity

Ton/ship - m3/ship

105000

80200

b, c

Base capacity

m3/silo

5000

2272

d

Base capital

M€

0.386

0.827

d

O&M

%

3

3

d

Storage characteristics

d Economic lifetime y 20 20 a The shipping costs are based on the figures provide in studies of Hamelinck et al. (Hamelinck et al. 2003; Hamelinck et al. 2005b) and own calculations. The shipping distance assumed in this study is the distance Maputo-Rotterdam (13355 km) and it is assumed that carriers will return empty. The port of Rotterdam is selected as it is one of the main ports of Europe and because Europe is expected to be an important market for bioethanol and (torrefied) pellets. It is assumed that all Mozambican harbours are able to facilitate large ocean

9

Considering a density of 0.6 and 0.8 ton/m3 for wood and torrefied pellets respectively, and a truck mass capacity of 40 tons, the transported volume is 67 and 50 m3, which is well below the assumed volume capacity of 130 m3. The density of ethanol is 0.789 and the assumed truck mass capacity is 25, resulting in a transported volume of 31 which is below the max capacity of 33 m3 (see Table 5.6). 177

SHADES OF GREEN bulk carriages and that they have sufficient loading capacity. The cost of shipping are lower than the charter prices reported by international traders (GF Energy 2010) and feasibility studies (Savcor 2006; Econergy 2008) of 40-57 €/ton pellets and 55 -68 €/m3 ethanol. b (Hamelinck et al. 2005b) c (Hamelinck et al. 2003) d (Hamelinck 2004)

The study of Junginger (2005) illustrated that no learning related cost reductions were identified in the transport stage in bioenergy supply chains in Sweden. However, in developing countries such as Mozambique, it is expected (and recorded) that road infrastructure is progressively being improved which is likely to reduce transport costs. Several rehabilitation schemes are implemented at present. On the other hand, the diesel subsidy which was long time provided by the Government of Mozambique ended in 2010. It is expected that global diesel prices will increase in the long term and therefore transportation costs will increase. In addition, toll charges are progressively implemented for new parts of infrastructure which also contributes to the cost of transport. It is not clear to what extend these counteracting forces level each other out. In addition, the improvement of road infrastructure is hard to grasp temporally and spatially explicitly, as it is uncertain in what time frame which stretches of road infrastructure are improved to which condition and if and where in what timeframe which types of roads are established. For these reasons, it is assumed that the differentiation in transport cost per tonkm per road type remain constant over time.

5.3 Results 5.3.1 Feedstock production costs The feedstock production costs decrease when yield levels per hectare increase as many of the cost items are per hectare costs. Figure 5.3, provides the cost breakdown for eucalyptus cultivation for increasing agro-ecological suitability in 2010 and 2030 and Figure 5.4 provides the cost breakdowns for sugarcane in these years. The total cost and the contribution of each cost item changes with increasing suitability and changes over time. For eucalyptus, the cost of land clearing and land preparation is significantly, but the contribution to the costs per ton eucalyptus reduces for more productive land. For sugarcane, the contributions of the costs for irrigation instalment and irrigating are the most profound. The costs for fertilizers and harvesting are cost ‘per ton feedstock produced’ and are assumed to be equal for all agro-ecological suitabilities and remain constant over time. However, the contribution of ‘per ton feedstock costs’ to the total costs increases for better suitabilities, as the relative contribution of ‘per hectare’ costs decreases.

178

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

Figure 5.3: Cost break down of eucalyptus cultivation for agro-ecological suitability levels in 2010 and 2030. Cultivation costs decreases by higher suitability levels as several cost items are per hectare and decrease per ton of product with higher yields. As the maximum yield level is higher in 2030, the cultivation costs per ton product are lower compared to 2010 for the same suitability level. As several key cost items (fertilizer application, harvest and extraction) are related to yield (and not to hectares) the cost decrease is asymptotic (and not linear).

Figure 5.4: Cost break down of sugarcane cultivation for several yield levels. Cultivation costs decreases by higher yield levels as several cost items are per hectare. As the maximum yield level is higher in 2030, the cultivation costs per ton product are lower compared to 2010 for the same suitability level. As several key cost items (fertilizer application, harvest and loading) are related to yield (and not to hectares) the cost decrease is asymptotic (and not linear).

In order to assess the effect of changes in individual cost items on the total production cost, a sensitivity assessment was performed. The sensitivity of the production cost for changes in individual cost items is specific for each suitability class and management factor and is therefore location and time specific. In Figure 5.5, the sensitivity of total cultivation cost for changes in individual cost items is depicted for a 75% suitability in 2010 which corresponds with a yield level of 17 odt/ha/yr for eucalyptus and 105 TC/ha/yr for sugarcane. The cultivation cost of eucalyptus are most sensitive to changes in costs of land clearing, inputs and harvesting and extraction, as these cost items contribute the most to the total cost. The cultivation costs of sugarcane are most sensitive for changes in the annual irrigation costs and to a lesser extent to harvesting cost and the cost of irrigation 179

SHADES OF GREEN

installation. Although the costs for land clearing are assumed to be equal for eucalyptus and sugarcane, changes in the cost for land clearing have very different effects on the total cost for eucalyptus and sugarcane. In addition, while fertilizer costs per hectare are much higher for sugarcane than for eucalyptus, the cultivation costs of eucalyptus are more sensitive to changes in fertilizer costs, as the relative contribution of fertilizer costs to the total cultivation costs are higher for eucalyptus than for sugarcane. The sensitivity of cultivation cost for higher yields would be lower for the ‘per hectare costs’ such as land clearing and irrigation, and relatively higher for the ‘per volume’ cost such as cost for fertilizers and harvesting.

Figure 5.5: Sensitivity analysis of feedstock cultivation costs. The sensitivity of the total cost of eucalyptus cultivation (left) is assessed for a suitability of 75% in 2010 which corresponds with as yield level of 17 odt/ha/yr. The sensitivity of the total cost of sugarcane cultivation (left) is assessed for a suitability of 75% in 2010 which corresponds with as yield level of 105 TC/ha/yr.

Figure 5.6 provides the development in the spatial distribution of potential production and costs of eucalyptus feedstock for two scenarios up to 2030. Figure 5.7 provides the maps of potential production and costs for sugarcane. The analysis on land availability and the results are presented in van der Hilst et al. (2011). As a result of technological learning, yield per hectare increase over time and therefore costs per unit produced feedstock decrease over time. In the progressive scenario, more land becomes available over time and the land that becomes available has a higher suitability on average. Therefore, there is a larger potential at lower cost available in the progressive scenario. For eucalyptus, the areas where feedstock can be produced below 4 €/odt increase over time and expand rapidly in the progressive scenario. The low production cost areas are located in the central south, the north east and the central part of Mozambique. Since eucalyptus has lower crop growing requirements, it is less susceptible for sub optimal growing conditions. Therefore, many of the available areas are suitable for relatively low cost eucalyptus production. For sugar cane the results are quite different. As only a few areas are considered to be suitable for commercial sugarcane growing, and cultivation cost increase significantly for decreasing suitabilities, the cost of sugarcane are very high in most areas. Only some areas in the central south, north east and central part of Mozambique have 180

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

suitable agro-ecological conditions for sugar cane cultivation. In these areas the cost of sugar cane could decrease to 13 €/TC. In the progressive scenario the areas where sugar cane cultivation is feasible increase significantly. In both scenarios, a large part of the available land is located in the central south and the eastern part (Tete province) of Mozambique. However, feedstock production costs are high in these areas due to marginal agro-ecological conditions. This is especially true for sugar cane that has high crop requirements and requires large investments per hectare.

181

SHADES OF GREEN

Figure 5.6: Spatial distribution of costs of eucalyptus feedstock production (€/odt) for the Business as Usual scenario (upper three maps) and the progressive scenario (bottom three maps).

182

Spatiotemporal cost supply curves for bioenergy production in Mozambique 5.

Figure 5.7: Spatial distribution of costs of sugarcane feedstock production (€/TC) for the Business as Usual scenario (upper three maps) and the progressive scenario (bottom three maps).

183

SHADES OF GREEN

5.3.2 Conversion costs The costs of pre treatment (eucalyptus to pellets) and conversion (sugarcane to ethanol) are calculated based on the input data provided in Table 5.4. The costs per GJ output are depicted in Figure 5.8. The production of torrefied pellets is expected to be slightly more expensive (2.70 €/GJ or 58 €/ton TOPs) than the production of wood pellets (2.20 €/GJ or 42 €/ton wood pellets) when only the pre treatment step is taken into account. However, the benefits of the torrefied pellets are found in transport and long distance shipping (see section 5.3.3) and the reduced investment costs for co-firing biomass in coal fired power plants. The costs of conversion of sugar cane to ethanol are expected to decrease from 3 3 6.20 €/GJ (130 €/m ) in 2010 to 4.70 €/GJ (100 €/m ) in 2030 and is mainly caused by an increased electricity output. The cost of conversion of sugarcane to ethanol cannot be compared directly to the cost of pellet production of eucalyptus because sugarcane to ethanol concerns a conversion step into a final product whereas the production from eucalyptus to (torrefied) pellets is only a pre treatment step. In addition, the costs for feedstock are not included in these cost calculations.

Figure 5.8: Cost of conversion from sugarcane to ethanol and cost for pre treatment of eucalyptus to wood pellets and torrefied pellets in 2010 and 2030 (excluding feedstock costs). The total ethanol production costs are reduced due to the revenues of electricity production.

5.3.3 Transportation costs Primary transport The cost of primary transport depends on the required input of the conversion plant, the distribution of the available land for bioenergy crops, and the yield level. In Figure 5.9 this relation is depicted for eucalyptus and sugarcane for the progressive scenario in 2010. As for eucalyptus the selected input requirement of the pre-treatment plant is fairly small, the transport distances are relatively short. In the progressive scenario, both the yield and 184

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

the land availability increase over time while the input requirements remain constant. Therefore, the average local transport distances of eucalyptus decrease towards 2030 from 5.8 to 5.3 km in the 20% most suitable areas and from 15.8 to 14.4 km in the 20% least suitable areas in the BAU scenario. In the progressive scenario the decrease in local transport cost is more profound: from 5.7 to 4.1 km in the 20% most suitable areas and from 15.6 to 9.7 km in the 20% least suitable areas. For sugarcane, the biomass input requirements of the ethanol plant are fairly large and increase over time due to the assumed larger plant capacities. In the BAU the local transport distance increases slightly due to a lower land availability from 16.5 to 16.6 km in the 20% suitable areas and from 71.8 to 77.4 km in the 20% least suitable areas. In the progressive scenario, the transport distance decreases from 14.6 to 12.7 in the 20% most suitable areas and 63.5 to 56.6 km in the 20% least suitable areas.

Figure 5.9: Relationship between land availability, yield and transport costs for eucalyptus (left) and sugarcane (right) for the progressive scenario in 2010.

It should be noted that average transport distances are not very representative as plants will only be established at locations with a high land availability and high yields. In order to assess which regions are most favourable from that perspective, the transport costs are depicted spatially explicitly in Figure 5.16 and Figure 5.17 the Appendix (Section 5.7). These figures show that for eucalyptus most available areas have low local transport cost due to the relatively small plant capacities. Only in areas with a low density of biomass (isolated patches of available land and/or marginal conditions), the cost of primary transport increase substantially. The variations in transport costs are much bigger for sugarcane. Especially in the eastern part (particularly Tete province) where the suitability for sugarcane cultivation is low, and in the isolated available areas, the costs of primary transport are high. In the progressive scenario the areas where primary transport is relatively cheap, increase over time.

185

SHADES OF GREEN

Secondary transportation The cost of secondary transport (pellets and ethanol from the plant to the harbour), depend on the distance of the production site to the nearest harbour and the availability and the condition of the road network. In Figure 5.10 the cost per ton goods is depicted spatially explicitly in relation to the road infrastructure and the location of the harbours. As all harbours are located at the Indian Ocean in the East, the transportation cost increases going westward. In the central south and central north the transportation costs are relatively high due to the low density of road infrastructure and the absence of (paved) primary roads. In the progressive scenario, more land becomes available in the coastal area in the North. Therefore, the areas with relative low cost for secondary transport increase over time in this scenario.

5.3.4 Total supply cost to international market The total supply costs for (torrefied) pellets and ethanol include the cost of feedstock production, primary transport of the feedstock from the field to the plant, the conversion or pre-treatment costs, the transport from the plant to the harbour, the storage and the international shipping costs. In Figure 5.11, the total costs for ethanol and (torrefied) pellets are depicted for 2010, 2020 and 2030 for the business as usual and the progressive scenario. The average costs represent the average costs considering all land available for energy crops for that specific year for that specific scenario. The average total costs provide little information about the economic performance of potential bioenergy supply chains, as they include also the isolated, remote low productive areas for which the cost are extremely high and that would never be included for commercial production. For that reason, also the total cost considering the 20 % most suitable areas are is included as these areas will be selected first for commercial production. For eucalyptus it is assumed that in 2010 wood pellets are produced and in 2020 and 2030 torrefied pellets are produced. As the torrefaction technology is expected to be more expensive than the production of wood pellets, the conversion costs increase over time. However, due to the higher mass and energy density of the torrefied pellets compared to wood pellets, the cost for secondary transport, storage and shipping are lower. The contribution of primary transport to the total costs is very small, due to the assumed small plant capacities and related input requirements. However, in the business as usual scenario, the average cost of primary transport increase due to a declining biomass density related to a decreasing land availability. The total cost of the sugarcane ethanol supply chain, are dominated by the cost of feedstock production.

186

Spatiotemporal cost supply curves for bioenergy production in Mozambique 5.

Figure 5.10: Spatial variability in transportation costs of the end products, (torrefied) pellets and ethanol, in €/ton in 2010, 2020 and 2030 for the Business as Usual scenario (upper three maps) and the progressive scenario (bottom three maps).

187

SHADES OF GREEN

In the sensitivity of the total cost of (torrefied) pellets and ethanol for changes in the cost of the individual stages of the supply chain are depicted for the eucalyptus pellet and the sugarcane ethanol chain. The sensitivity analysis provided here, is the sensitivity of the average cost of the 20% most suitable areas in the progressive scenario in 2030. However, the sensitivity of the total cost for changes in cost items is highly spatially and time dependent. The sensitivity analysis shown in Figure 5.12 is therefore only an example. In Table 5.7 the ranges of the costs of the individual stages used in the sensitivity analysis are depicted. The footnotes of Table 5.7 provide justifications of the selected ranges. Table 5.7: Ranges for sensitivity analysis of total costs of supply chain for eucalyptus and sugarcane in 2030 in the Progressive scenario. Stage of supply chain Feedstock a st

1 Transport

b

Conversion c nd

2 Transport

d

Eucalyptus

Sugarcane

Unit

Normal

Range

Unit

Normal

Range

€/odt

35

22 - 124

€/TC

14.57

6.76 – 132

€/ton

0.62

0.03 – 2.56

€/ton

1.91

0.10 – 7.75

€/ GJ

2.70

1.35 – 5.40

€/ GJ

4.85

2.43 – 9.70

€/ ton

6.96

0 – 63

€/ ton

6.96

0 – 63

Storage e

€/ ton

1.36

0 – 14.29

€/ m3

0.58

0 – 5.77

Shipping f

€/ ton

19.30

0 – 57

€/ m3

15.30

0 – 68

a

The low end of the range of feedstock production cost of eucalyptus is based on the for the best suitable soils in Mozambique and excluding costs for land clearing (excluding land clearing). The low end of the sugarcane production costs is based on the lowest projected feedstock production cost of Brazil in 2020 (van den Wall Bake et al. 2009). The high end of the range for both eucalyptus and sugarcane is based on the production cost of the least suitable soils (10%) in Mozambique. b The low end of the range is based on a area with 100% availability and 100% suitability and a good road network. The high end is based on a area with a low availability (10%) and low suitability (10%) and no proper road network. c There is a high uncertainty in the development of technological learning and related cost reductions. For that reason a range of 50-200% of the estimated conversion cost is included. d The low end of the range is for the situation that the plant is located in the harbour, so no transport is required. The high end of the range is based on the transport distance from the least accessible location to the harbour. e The low end of the range is for the situation that no storage is required (production meets downstream logistics). The high end of the range is for the situation that there is only 1 pellet /ethanol plant in Mozambique (capacity see Table 5.4) and that the pellets/ethanol is only shipped when the full capacity of the ship (see Table 5.6) is used. Storage costs are depicted in Table 5.6. f The low end of the range is applicable when there is no international shipping required (domestic use). The high end represent the figures reported by international traders (GF Energy 2010) and feasibility studies (Savcor 2006; Econergy 2008).

The sensitivity analysis shows that the cost of the pellet chain is predominantly sensitive for variations in the feedstock and conversion costs but also for changes in the shipping and secondary transportation. The total costs are little influenced by the cost for primary 188

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

transport (because of the small capacity of the pre treatment plant) and the cost for storage. For sugarcane, the total cost are predominantly susceptible for changes in feedstock and conversion cost. In contrast to the eucalyptus supply chain, the total cost of the sugarcane ethanol supply chain are quite susceptible for changes in secondary transport costs due to the relatively large input requirements of the plant, the low energy density of the feedstock, and the relatively low conversion efficiency. In order to assess which areas are most favourable for bioenergy production considering the total cost for the entire supply chain up to delivery in an international harbour, the cost are depicted spatially explicitly for both scenarios up to 2030 in Figure 5.13 and Figure 5.14. The Figures show that the cost range of the total supply chain is very large. The costs of the eucalyptus pellet chain are very high in the central south and the north east (Tete province). These areas have a low productivity and therefore high feedstock production costs, and are relatively remote which results in high cost for secondary transport. In the central part of Mozambique there is not much land available for bioenergy production. However, the areas that are available are relatively favourable areas due to high productivity and proper accessibility by road network. The north eastern part of Mozambique has the most favourable conditions as these areas are high productive and are situated close to the harbour of Nacala and Pemba. The spatial pattern of production cost of sugarcane ethanol appears to be quite similar to the pattern of wood pellet production costs. The spatial variation is however even more related to the spatial variation in feedstock production costs due to the high contribution of feedstock cost to the total costs. In addition, the spatial variation in total cost is much larger compared to eucalyptus pellet costs. The areas with low supply chain cost are far more scarce than for eucalyptus pellets as only a few areas are very suitable for sugarcane cultivation.

Figure 5.11: Total cost of bioenergy supply chains of (torrefied) pellets (left) and ethanol (right) for 2010, 2020 and 2030 for the Business as Usual and the progressive scenario, considering the average of all available land (‘Average’) and the 20% most suitable areas (‘Suitable’).

189

SHADES OF GREEN

Figure 5.12: Sensitivity of total supply cost of torrefied pellets (left) and ethanol (right) in 2030 in the progressive scenario for changes in cost of supply chain stages. The sensitivity of the total supply cost for the 20% most suitable area is calculated.

190

Spatiotemporal cost supply curves for bioenergy production in Mozambique 5.

Figure 5.13: Total cost of supply chain of eucalyptus wood pellets (2010) and torrefied pellets (2020 and 2030) for the business as usual and the progressive scenario. Total costs include cost of feedstock cultivation, primary transport, pre treatment, transportation form plant to harbour, storage and international shipping.

191

SHADES OF GREEN

Figure 5.14: Total cost of supply chain of sugarcane ethanol in 2010, 2020 and 2030 for the business as usual and the progressive scenario. Total costs include cost of feedstock cultivation, primary transport, conversion, transportation form plant to harbour, storage and international shipping.

192

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

Figure 5.15: Cost supply curves of (torrefied) pellets (left) and sugarcane ethanol (right) for 2010, 2020 and 2030 in the Business As Usual and the progressive scenario. The cost supply curve is a ranking of the supply potential according to the total supply costs.

Based on the time and spatially explicit calculations, dynamic cost supply curves can be constructed for torrefied pellets and sugarcane ethanol supply chains (see Figure 5.15).The cost supply curves rank the potential supply according to the total cost of the supply chain which includes feedstock production, primary transport, pretreatment/conversion, secondary transport, storage and international shipping. The solid lines represent the cost supply curves for the Business as usual scenario for 2010, 2020 and 2030. The dashed lines represent the cost supply curves for the progressive scenario. The supply curves of both bioenergy chains show that in both scenarios, the costs decrease over time. However, the low-cost production potential is much bigger in the progressive scenario compared to the business as usual scenario. For eucalyptus pellets the total potential is quite large (3200 PJ in 2030 in the progressive scenario), especially compared to the potential of sugarcane ethanol (866 PJ in 2030 in the progressive scenario). This is due to two main reasons: First, sugarcane is already converted to ethanol in which energy is lost, whereas pellets are still about to be converted to power and heat. And second, much more land is suitable for eucalyptus cultivation than for sugarcane cultivation. Like all energy carriers, the markets for pellets and ethanol are quite volatile. The study of Sikkema et al.(2011a) on wood pellet trade in Europe shows a CIF price fluctuation between 6 and 8 €/GJ in Europe in the timeframe 2008-2010. Assuming the lower end of this price range, 20 PJ in 2010 and 70 PJ in 2030 in the BAU scenario and 25 PJ in 2010 and 550 PJ in 2030 in the progressive scenario could be exported to Europe below this price level. Considering the higher end of the price range, 530 PJ in 2010 and 880 PJ in 2030 in the BAU scenario and 640 in 2010 and 2520 PJ in the Progressive scenario could be exported to Europe for a cost level below the market price. For comparison: in 2009 an equivalent of 30 PJ pellets was imported by Europe (EU-27) (Sikkema et al. 2011a) .The 3 prices of FOB T2 (duty paid) ethanol in Rotterdam fluctuated between 555 and 665 €/m in 2010 (Platts 2010). As Mozambique belongs to the least developed countries (LDC), 193

SHADES OF GREEN

duty free access is granted by the EC by means of the everything but arms (EBA) regulation (European Commission 2001). In the BAU scenario, it is possible to export 100 PJ ethanol in 2010, and 135 PJ ethanol in 2030 from Mozambique to Europe for a cost level below the lower end of this price range. In the progressive scenario 130 PJ in 2010 and 405 PJ in 2030 could be exported below this threshold. Considering the higher end of the price range, 140 PJ in 2010 and 165 PJ in 2030 in the BAU scenario and 165 PJ in 2010 and 515 PJ in 2030 in the progressive scenario could be exported to Europe. For comparison: in 2010, 50 PJ ethanol was traded internationally worldwide (OECD and FAO 2011).

5.4 Discussion This study has assessed the potential developments in potentials and cost of bioenergy supply chains in Mozambique spatially explicitly in the timeframe 2010-2030. Assessing the potential future of bioenergy in Mozambique comes with numerous uncertainties in the developments in several different areas such as agriculture, technology, infrastructure, commodity markets. However, the aim of this study is especially to demonstrate how cost and supply developments could be assessed temporarily and spatially explicitly, and not to provide all possible future outcomes of cost supply potentials of Mozambique. 2

The development in potential cost supply of bioenergy is calculated on a 1km grid size level. For this reason, landscape characteristics that play a role on a more detailed level that could affect cost and supply are not taken into account (e.g. a grid cell is either available for energy crops or not, but not partly; detours required due to small river streams; elevations or gaps are not considered). The suitability maps of FAO and IIASA used to assess the spatial variation in yield level of eucalyptus and sugarcane, have a relatively coarse resolution (5arcmin) compared to all other spatial data inputs. As there are no suitability maps available specifically for eucalyptus a general suitability map combined for crop and pasture is used as a proxy indicator. For sugarcane, the suitability map for irrigated and high input management is used in this study. Although, it takes into account whether climate, soils and terrain allow for irrigated crop cultivation, it does not consider the water availability within a watershed (Fischer et al. 2000). More detailed and crop specific suitability maps including additional factors affecting the suitability, would contribute to more accurate estimations of attainable yield levels. Cost data of eucalyptus and sugarcane cultivation, transport, pre treatment, conversion, and shipping are derived from extensive literature review and field work. The results of the study show that the cost ranges of eucalyptus feedstock production are 1.81-2.40 €/GJ in 2010 and 1.60-2.04 in 2030 in suitable areas in Mozambique. Costs of eucalyptus in 194

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

Brazil reported in literature vary substantially between 0.5 €/GJ(standing stock) (Walter et al. 2006) and 2.3 €/GJ (de Wit et al. 2011b). The production costs of sugarcane cultivation of 14-22 €/TC in 2010 and 13-21 €/TC in 2030 found for the 20% most suitable areas in Mozambique, are within the range of sugarcane production costs reported in literature of 13-20 €/TC in the US (Pimentel et al. 2008; Salassi and Deliberto 2010), 9-30 €/TC in Brazil (Marques et al. 2009; van den Wall Bake et al. 2009; Xavier et al. 2009; Crago et al. 2010; Smeets and Faaij 2010) and 14-20 €/TC in Mozambique (Econergy 2008; Jelsma 2010). Reported feedstock production costs vary widely due to different agro-ecological conditions, different plantation designs, different management practices and differences in local costs of inputs. The variations in estimations of production costs of several studies are also caused by differences in system boundaries, and applied exchange rates and discount factors. In this study, it is assumed that feedstock production costs are reduced over time due to yield increases. However, some studies show that cost of feedstock production do not necessarily decrease over time and could even increase (Breux and Salassi 2005; Salassi and Breux 2006; Salassi and Deliberto 2007; Salassi and Deliberto 2008; Salassi and Deliberto 2009; Açucareira de Xinavane SA 2010; SA Cane Growers association 2010; Salassi and Deliberto 2010; Salassi and Deliberto 2011). In recent years, the feedstock costs have been greatly affected by increased costs of equipment, diesel, fertilizers and agrochemicals (BNDES and CGEE 2008). It is expected that these costs continue to increase in line with historical global trends (Mitchell 2008; ESR and USDA 2011). Currently, the wages are very low in Mozambique compared to other countries (Hoogwijk et al. 2009). As the GDP is projected to continue to grow (6.6 %/y) [47], it is expected that labour wages will go up. As in Mozambique the pressure on land is expected to increase due to population growth (UNDP 2008) and increased interest from investors, land is likely to become more expensive. In addition, apart for the cost for the land lease, there are also cost associated with compensation for communities in the vicinity of the project which could go up when the higher pressure on lands results in that more people get affected by the implementation of a large scale project (these costs are not included in this study). Although it is expected that by means of more efficient management (improved planting material, reduced pest management, more advanced equipment) these cost rising factors could be counteracted, future cost figures could be underestimated. However, this cost increases will affect bioenergy production cost worldwide. Simultaneously, it is expected that fossil fuel prices will increase, especially when CO2 prices are included, which could make bioenergy more competitive in the future. Reported cost estimates of ethanol and wood pellet production vary widely due to large differences in plant lay out and scale. Costs for torrefied pellet production are hard to 195

SHADES OF GREEN

estimate as no commercial plant is in operation yet. The revenues from electricity in the ethanol production process estimated in this study are relatively low compared to other countries (such as Brazil). This is related to the electricity price in Mozambique(0.036 €/kWh) (CPI 2009) which is reasonably low due to the high potential supply of hydropower plant of the Cahora Bassa in Tete province. However, in areas that are not equipped with an electricity grid, the cost of electricity could be much higher. Although the potential revenues from electricity are higher in those areas, in absence of a grid, demand is still low. However, due to investments and electrification projects, the national grid is expanding rapidly. The cost of inputs for pre treatment and conversion (energy, materials, labour) could become cheaper due to better logistical chains and higher domestically available skilled labour (less imported labour required), but conversion could also become more expensive due to higher energy cost (and related transportation costs) and increased labour wages due to higher economic development. Torrefied pellet have higher conversion costs but have more favourable characteristics for handling and storage. This may result in a premium price for TOPs but this is not considered here. In this study it is assumed that there are no economies of scale beyond a scale of 40MWth(in), in line with the study of Uslu et al. (2008). However, depending on the type of technology and technological developments, economies of scale could be achieved which would result in lower conversion costs. The costs of transport are relatively uncertain as broad ranges are mentioned in literature and in interviews with stakeholders. In this study a cost range between 0.02 €2010 for paved primary roads in good condition and 0.15 €2010 for off-road transport is assumed. This range is in line with most reported costs in literature (Meeuws 2004; GDS 2005; Tostao and Brorsen 2005; Econergy 2008; USAID 2008), it is however lower than the cost of 0.40 €2010 per tonkm for liquids and 0.59 €2010 per tonkm for wood reported by some stakeholders interviewed for the studies of (Econergy 2008) and (GDS 2005); and it is higher than the range of 0.026-0.043 €2010 tonkm used in the study of Batidzirai et al. (Batidzirai et al. 2006). The allocation of variations in transport cost (€/tonkm) to the several road types and conditions is somewhat artificial. However, it is assumed to be more realistic than applying one cost level for the entire road network. The transportation costs and spatial variation in them could be quite different when improvements in the road infrastructure and transport by railroad and water ways would have been included. However, truck transport is currently preferred due to the low reliability and capacity of railroads (Meeuws 2004) and because waterways are currently not used for transport. However, ongoing rehabilitation of roads and railways is scheduled and the use of waterways for transport (e.g. Zambezi River for the transport of coal from Tete) is under investigation and discussion. Improvement of the road infrastructure and the use of waterways and railways would require large investments but could reduce transportation 196

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

cost significantly in the long run as tonkm railway transport cost are currently estimated at 0.024-0.065 €/tonkm (Meeuws 2004) and domestic shipping is estimated at 0.027 €/tonkm and could be reduced by means of up scaling and optimization of logistical planning. None of the harbours included in this study is equipped for large scale biomass export yet. All harbours would require considerable investments in order to facilitate large shipping quantities. The investments and the growth potentials vary for each harbour and are not included in this study.

5.5 Conclusions In this study, the development of potentials and costs of bioenergy supply chains in Mozambique in the timeframe 2010-2030 is assessed spatially explicitly. The results show that there is currently a large potential for bioenergy production in Mozambique and that this potential could increase significantly towards 2030 in the Progressive scenario. Considering the upper limit of current market prices of 7.88 €/GJ for pellets, the economic viable pellet production potential increase from ± 500 to ±900 PJ in the BAU scenario and from ±600 to ±2500 PJ in the Progressive scenario (which corresponds to ±20 % of OECD Europe’s current coal consumption ±13 EJ (IEA and OECD 2011)). Considering the higher end of the ethanol price range of 2010, the economically viable ethanol supply potential is ± 150 PJ in the BAU scenario and increases from ± 150 to ± 500 PJ in the progressive scenario (which corresponds to 7% of Europe’s current gasoline consumption of ± 6 EJ (European Commission 2006)). Optimisations in the supply chains could be achieved by improved breeding for higher yield levels and reduced management costs of feedstock production; technological learning and scale increase of the pre-treatment and conversion technology; decentralised pre-treatment and biorefinery; and improved of infrastructure including road, rail, waterways and pipelines. Further research on these optimisations of the bioenergy supply chain is required. The results show a large spatial variation in supply chain costs which is the result of spatial variation in feedstock production costs, primarily transport costs and secondary transport costs. These costs have all their own spatial pattern. Some productive locations enabling low feedstock production costs are very isolated and/or very remote (e.g. in the North West of Niassa); and some areas are very well accessible but not productive (e.g. central east Cabo Delgado). The areas where low supply chain production cost can be achieved are those areas that are productive, well surrounded by more productive land and easily accessible. Areas meeting these conditions are scattered in the central south, central part of Mozambique and the central-east of the North. As feedstock cost is the main contributor to the total cost of the ethanol supply chain, the spatial heterogeneity of agroecological suitability has the strongest relationship with the spatial pattern of the total 197

SHADES OF GREEN

cost of the sugarcane ethanol supply chain. The spatial pattern of the total supply costs of eucalyptus pellets is also mostly determined by the spatial pattern of the agro-ecological suitability but also by the spatial pattern of the cost of secondary transport. This study demonstrates an approach which enables the assessment of the development of bioenergy potentials and costs over time in a spatially explicit way. This results not only in a more accurate estimation of actual potentials and how they will develop over time, but provide also information on which areas are most viable for bioenergy production taken into account all cost of the whole supply chain. This means a large step forward in the assessment of bioenergy potentials. As environmental and socio-economic impacts of bioenergy supply chains are highly related to the physical and socio-economic context of the location of production, a spatially explicit assessment of bioenergy production potential is a suited approach for the assessment of the sustainability of bioenergy chains.

5.6 Acknowledgements This study is part of the Climate Changes Spatial Planning Programme and has been funded by the Dutch government, the European Commission and Shell. In addition, it has been partly funded by the Biorenewable Resources Platform, SASOL and UNEP. The authors gratefully acknowledge governmental and non-governmental institutions, and agricultural, forestry and biofuel companies in Mozambique for their contribution to the data gathering and for sharing their knowledge and expertise. In addition, the authors would like to thank Professor Johan Sanders for his contribution to this study, Janske van Eijck for her contribution to the field work; Martin Junginger for sharing his knowledge on technological learning and Judith Verstegen for commenting on the draft paper.

198

5.

Spatiotemporal cost supply curves for bioenergy production in Mozambique

5.7 Appendix: Additional results primary transport costs In Figure 1 and 2, the cost of primary transport from field to plant are depicted for Eucalyptus and sugar cane for 2010, 2020 and 2030 for the Business as usual and the progressive scenario. Figures 1 show that for eucalyptus most available areas have low local transport cost due to the relatively small plant capacities. Only in areas with a low density of biomass (isolated patches of available land and/or marginal conditions), the cost of primary transport increase substantially. The variations in transport costs are much bigger for sugarcane (see Figure 2). Especially in the eastern part (particularly Tete province) where the suitability for sugarcane cultivation is low, and in the isolated available areas, the costs of primary transport are high. In the progressive scenario the areas where primary transport is relatively cheap, increase over time.

199

SHADES OF GREEN

Figure 5.16: Spatial variation in costs of primary transport of eucalyptus (€/ton) for the Business as Usual scenario (upper three maps) and the progressive scenario (bottom three maps).

200

Spatiotemporal cost supply curves for bioenergy production in Mozambique 5.

Figure 5.17: Spatial variation in costs of primary transport of sugarcane (€/TC) for the Business as Usual scenario (upper three maps) and the progressive scenario (bottom three maps).

201

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

6

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine F. van der Hilst, J.A. Verstegen, T. Zheliezna, O. Drozdova, A.P.C. Faaij

203

SHADES OF GREEN

ABSTRACT This study analyses how the bioenergy potential and total greenhouse gas (GHG) balances in Ukraine develop over time, taking into account changes in emissions of total agricultural land use. The development in land requirements for food and feed production is analysed spatially on an annual basis making use of the PCRaster Land Use Change (PLUC) model. Two scenarios for the period 2010-2030 have been assessed: a Business As Usual scenario (BAU), in which current trends in productivity are continued; and a progressive scenario, which projects a convergence of yield levels for Ukraine with West European countries. In the progressive scenario, 32.1 Mha land could become available for energy crop production by 2030. The projected land use developments serve as input for the spatiotemporal GHG balance, which includes CO2, N2O and CH4 emissions related to changes in management and land use, as well as the abatement of GHG emissions by replacing fossil fuels by assumed bioethanol production from wheat and switchgrass. The spatiotemporal 2 GHG module produces spatially explicit maps (1 km resolution) of the individual GHG emissions on an annual basis. The results show that a total cumulative GHG balance of -0.8 GT CO2-eq for wheat and -3.8 GT CO2-eq for switchgrass could be achieved in 2030 in the progressive scenario. When measures are taken to reduce agricultural CO2 and N2O emissions, the cumulative GHG balance could even increase to -2.6 GT CO2-eq for wheat and -5.0 GT CO2-eq for switchgrass by 2030. When the available land is used for the re-growth of natural vegetation, a considerable amount of carbon will be accumulated in the form of biomass and soil organic carbon. This could reach -4.3 GT CO2-eq in 2030. However, this carbon sequestration can only be obtained once in contrast to the abatement of bioenergy crops. The spatiotemporal GHG module in conjunction with the PLUC model allows for spatially explicit and dynamic modelling of total GHG emissions that result from land use and management changes related to the implementation of bioenergy crop production. 204

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

6.1 Introduction Competition between food, feed and fuels, and therefore indirect land use change (iLUC), could be avoided if the increased production of biomass for energy is balanced by improvements in agricultural management (Dornburg et al. 2010; Wicke et al. 2012). Therefore the amount of bioenergy that can be produced without expanding the total agricultural land use area depends on the rate of intensification of the agricultural sector and the suitability of the land that becomes available for energy crop production. However, both intensification of the agricultural sector and land use change can have adverse environmental impacts such as GHG emissions related to land conversions, and N2O emissions for amplified fertiliser use, which could diminish the GHG mitigation potential of the substitution of fossil fuels. Potential yield levels and environmental impacts of intensification of the agriculture and energy crop production are strongly related to the biophysical conditions of a region (van Dam et al. 2009a; 2009b; Van der Hilst et al. 2010; Beringer et al. 2011; Van der Hilst et al. 2012a); Therefore, it is important to assess both the potentials and the potential impacts of the intensification of the agricultural sector and the implementation of bioenergy crops spatially explicitly, taking the variability in land use, soil, and climate into account. Studies on global and European bioenergy potentials have indicated large technoeconomic production potentials for Eastern Europe and for Ukraine specifically (Smeets et al. 2007; de Wit and Faaij 2010; Fischer et al. 2010a; de Wit et al. 2011a). Ukraine is considered to be a promising region for bioenergy production because of favourable climate conditions, rich agricultural resources (high quality soil and land suitable for irrigation), access to abundant water resources, and the proximity to major foreign markets (Morton et al. 2005). The decreasing population, the stable dietary intake, and the low current agricultural productivity provide opportunities for decreasing the area required for food and feed production, and thereby for increasing the potential land available for bioenergy crop production. Several studies have assessed the biomass production potentials in Europe. However, only a few studies have done this spatially explicitly (Hellmann and Verburg 2008; de Wit and Faaij 2010; Fischer et al. 2010a; Fischer et al. 2010b). Only the studies of Fischer et al. (2010a; 2010b) and de Wit and Faaij (2010) 10 have included Ukraine in their assessment. However, as this study was done on a NUTS 2 level, little information is available on the exact amount and the location of available land for energy crops in Ukraine. Therefore in this study, the PCRaster Land Use change (PLUC) model (Van der Hilst et al. 2011; Verstegen et al. 2011) is used to assess the potential land

10

Nomenclature of Units for Territorial Statistics 205

SHADES OF GREEN

use change developments and related land availability for energy crops in Ukraine on a detailed spatial level. Some studies have integrally assessed the impacts of agricultural and use and land use changes on GHG emissions (Smith et al. 2000; Leip et al. 2008; Lesschen 2008; Smith et al. 2008). Popp et al. (2011) assessed the GHG emissions of bioenergy production including the co-emissions from agricultural intensification on a spatial level of 3 degrees (i.e. about 300km at the equator). De Wit et al. (2011c) assessed environmental impacts of integrating biofuels in the agricultural sector in Europe on a NUTS 2 level, taking into account the agricultural intensification. Due to the scope of these studies, the assessments were made on an aggregated spatial level, which does not recognise the spatial variation in emissions due to land use, soil type, temperature and precipitation, etc., while these differences in conditions result in very high variations of GHG emissions. Therefore, GHG emission should be assessed on a spatially detailed level taking into account variations in the biophysical factors (Horner et al. 2011). In this study, a dynamic model is developed to assess the developments in CO2, N2O and CH4 emissions temporally and spatially explicitly for the period 2010-2030, taking into account the emissions related to agricultural intensification and land use change. The emissions are differentiated for local conditions such as land use, climate, soil, and crop composition, making use of the best quality GIS data available. It is assumed that agricultural land that is abandoned could be used for the production of energy crops. Wheat and switchgrass are selected as typical first and second generation bioethanol crops respectively. As re- and afforestation could also offer high potentials for carbon sequestration (Kuemmerle et al. 2011), re-growth of natural vegetation is also considered as an alternative use of abandoned farmland.

6.2 Methods The conducted analysis consists of 4 main steps: 1. Projections are made for the development in domestic production of food, feed and livestock products. 2. Scenarios are constructed on the developments in agricultural productivity. 3. Land use changes are modelled spatially explicitly making use of the PCRaster land use change model (PLUC). 4. The GHG emissions due to changes in land use and management are analyzed spatially explicitly.

206

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

The GHG emission reduction potential of biofuel production in combination with agricultural intensification is compared to the current situation with low productive agriculture and to a future situation in which carbon sequestration due to re-growth of (semi-) natural vegetation occurs on abandoned agricultural land. The four methodological steps are described in detail below.

6.2.1 Future demand The demand for domestically produced food and feed is related to developments in population size, Gross Domestic Product (GDP), food intake per capita and self sufficiency ratio (SSR, i.e. the extent to which domestic supply meets domestic demand (FAO 2003b)). The future food and feed production requirements are based on the outlook provided by the FAO (2011). This outlook provides aggregated figures for the group ‘Other Eastern 11 European Countries’ for the period 2006-2050. The production growth rate in the period 2006-2030 for the aggregated Eastern European countries was differentiated in population related growth; dietary intake related growth and export related growth. These differentiated growth rates were applied for the period 2010-2030 starting with the current production levels of Ukraine derived from (FAO 2010a; State Statistics Service of Ukraine 2011) and accounting for the projected population growth in Ukraine for that period by the UNDP (2009) and for the convergence of dietary intake of Ukraine in cal/capita/day with current average European levels in 2030 derived from (FAO 2010a).

6.2.2 Scenarios During the Soviet period (until 1991), arable land in Ukraine was almost completely state owned and managed by large agricultural enterprises (kolkhoz and sovkoz) (Kuemmerle et al. 2006; de Wit et al. 2011a). Agricultural production saw modest growth that was stimulated by state investments, thereby implicitly subsidising the agricultural sector (de Wit et al. 2011a). After 1990, the shift from a socialist system to a market oriented economy resulted in major changes in land ownership and fragmentation of farm fields due to land reforms (Kuemmerle et al. 2006). During this period, agricultural production dropped dramatically due to the obstruction in the land market, the reallocation of agricultural capital, the abrupt abolishment of subsidies, liberalisation of prices, and the low purchasing power (Osborn and Trueblood 2002; de Wit et al. 2011a). For instance wheat yield levels decreased from 3.5 ton/ha in 1990 to below 2 ton/ha in 2000. Due to these developments, land abandonment was occurring at unprecedented rates and large areas of arable land were converted to grassland and forest (Kuemmerle et al. 2006). In addition, agricultural employment went down, which was partly absorbed by subsistence 11

The group ‘Other Eastern European Countries’ consist of Ukraine, Albania, Bosnia and Herzegovina, Croatia, Montenegro, Republic of Moldavia, Republic of Macedonia and Serbia. 207

SHADES OF GREEN

farming on small plots (World Bank and OECD 2004). In the period between 1990 and 2008, the use of fertilisers, pesticides and machinery decreased rapidly (FAO 2010a). Subsistence farming contributes to low overall agricultural productivity due to a lack of economies of scale, crop specialisation and access to machinery, inputs and markets. Over the last decade, agricultural production and exports have increased in Ukraine, although this is mainly attributed to more favourable weather conditions (de Wit et al. 2011a). Although several studies project a large bioenergy potential for Ukraine(Smeets et al. 2007; de Wit and Faaij 2010; Fischer et al. 2010a; Fischer et al. 2010b; de Wit et al. 2011a), the political, institutional and technical changes required to accomplish this, are significant. For this reason, a Business As Usual (BAU) and a progressive scenario (PROG) are used to explore the broad range of possible future developments. The BAU scenario represents a continuation of historic trends in agricultural productivity (after the dissolution of the Soviet Union). This is in line with the ‘low’ scenario of the Refuel analysis 12 for Central and Eastern European Countries (CEEC) (de Wit and Faaij 2010) and with the projections of the FAO. Trends in yield figures per crop for the period 1992-2009 are extrapolated towards 2030. This implies for instance that the average yield level of wheat in the period 1992-2009 of 2.7 ton/ha is maintained towards 2030. The progressive scenario represents a future with full implementation of agricultural and institutional reforms that enable (foreign) investment in the agricultural sector and the adoption of more advanced agricultural technologies and management practices resulting in a convergence of yield levels of Ukraine with West European Countries (WEC) by 2050. This implies that the yield level of wheat will increase from an average of 2.7 ton/ha to 6 2 ton/ha in 2030. This is in line with the ‘baseline’ scenario of the Refuel analysis . In Table 6.1 the key characteristics of the scenarios are described. In addition to the progressive scenario, a scenario with more emphasis on sustainable agriculture is developed. This scenario is similar to the development in the progressive scenario, but it is assumed that mitigation measures to reduce agricultural GHG emissions such as reduced tillage, increased carbon input, and fertiliser type improvements are implemented.

12 In the Refuel project the cost and supply potential for biomass resources were assessed for the EU27 and Ukraine for the period up to 2030. One of the key methodological steps was to determine the surplus agricultural land based on the domestic demand, self sufficiency ratio and agricultural productivity. Three scenarios were developed for agricultural productivity. For the CEEC (Central and Eastern European Countries) the low scenario assumed a continuation of current yield levels. The ‘baseline’ scenario assumed a convergence with WEC agricultural productivity levels in 2050 and the high scenario assumes this convergence to be in 2030.

208

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

Table 6.1: Main characteristics of the two scenarios related to land use change Characteristic

Current (2010)

Population a Diet a, b SSR b Farming practices c

45.4 Mpeople 3200 kcal/cap 1.14 Combination of: - household subsistence farming (1 - 2.5 ha) 38% of agricultural land. - private-individual farms (70 -80 ha) 8% - large enterprises (>1000ha) 54%. Low use of irrigation (<2% of arable land), fertilizers (<30 kg nutrients/ha), pesticides (<0.8 kg/ha), and machinery. Low crop yield (average 2.7 t/ha wheat), low cropping intensity (CI) 0.77.

Technology adoption a, d

Agricultural productivity e

Future (2030) BAU

PROG 40.5 Mpeople 3300 kcal/cap 1.29 Continuation of trend Abandonment of towards more subsistence farming, household and private increase in number and farms at the expense of size of private-individual large enterprises. farms and reform of large enterprises.

Little improvement in accessibility of inputs and machinery.

High adoption inputs and machinery; meets West European practices

Little development in yields and modest increase in cropping intensity (0.3% p.a)

High increase in crop yields (3.8% p.a.) and cropping intensity (1.2% p.a.) resulting in a CI of 1 in 2030. Increase in livestock numbers (same as BAU), shift towards high productive farms, full mechanisation and the use high quality fodder. Similar practices and productivity in livestock sector as in Western Europe by 2030.

Livestock sector a, f

Low livestock numbers Increase in livestock compared to historic levels. numbers and modest 66% of cattle, 55% of pigs, increase in 83% goats and 46% of poultry productivity. Due to is produced on household modest shift from small farms (State Statistics Service to large production of Ukraine 2011) which are farms. characterised by low number of animals, manual labour and low nutritious fodder which results in a low productivity (Bogovin 2001). Low carrying capacity pastures. Bioenergy No significant commercial Abandoned agricultural land is used for bioethanol implementation g bioenergy production. crops. a Figures are derived from FAOSTAT b Derived from FAO Agricultural outlook c Based on Kucher, (2007) and Lerman et al. (2006). Subsistence farmers that produce mainly for own consumption and supply limited proportions of their production to the market (20-40%). d Assumptions on future farming practices in progressive scenario are based on (de Wit and Faaij 2010; de Wit et al. 2011a) e The crop productivity is defined by the yield levels in ton/ha and the cropping intensity (area harvested/arable land). Low cropping intensities indicate large amounts of fallow /set aside land.

209

SHADES OF GREEN f

Development in livestock production efficiencies for eastern European and western European countries are derived from Bouwman et al. (2005). g Currently 1 million hectares of land is in used for rapeseed production in Ukraine. Although recent trends show a decrease (State Statistics Service of Ukraine 2011), in this study it is expected that rapeseed production will double by 2030 (FAO outlook) in line with current trends and announced policies. GHG emission avoidance by substituting biodiesel is not incorporated in this study.

6.2.3 Land use change model Land availability for energy crop production is modelled taking into account the land required for other land use functions such as nature conservation and food production. In order to spatially asses land use change dynamics, the PCRaster Land Use change (PLUC) model has been developed. The methods and the data inputs for the modelling of land availability for energy crop production are described in Van der Hilst et al. (2011). The technical characteristics of PLUC are described in Verstegen et al. (2011). The land use changes in the period 2010-2030 were modelled for each year on a grid with 2 a 1km cell size, by allocating land to a specific land use class based on the total demand for the products from this class (pasture, crops) and the location-specific suitability of for this land use class. The suitability of land for a specific land use is defined by a spatial weighted summation of a specific set of suitability factors (such as the agro-ecological suitability, population and cattle density, and accessibility (i.e. distance to main cities and infrastructure). Areas that are not suitable (such as steep slopes) or not allowed (such as conservation areas) to be converted to agricultural land, are excluded from expansion of agricultural land including energy crops. The allocation of land to the land use classes in order to meet the demand results in a new land use map for that time step (year). The modelling comprises a feedback loop: the map of the result of the allocation of time step t is an input for the allocation in time step t+1. Based on this land use map, land can be allocated to bioenergy crops or to the re-growth of natural vegetation. As it is a starting point that bioenergy crops should not compete with food and feed production, bioenergy crops are allocated last, when all other land use requirements for that time step are met. Agricultural land that becomes available in the excluded zones (e.g. steep slopes, conservation areas) is allocated to the re-growth of natural vegetation. The PLUC model is tailored for the Ukrainian context by applying country specific drivers and suitability factors for land use change. In some LUC models, a multiple regression model is applied to identify the driving forces and assess the influence of these factors on land use, e.g. in Veldkamp et al. (1996) and Verburg et al. (1999). Preferably, historical land use patterns and drivers are recorded for several points in time spread over a long time period in order to calibrate the suitability factors used for allocation. However, these data are not available for Ukraine. Moreover, extrapolation of regression analysis may 210

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

produce dubious results as historical driving forces for LUC may change or no longer be relevant (Veldkamp and Verburg 2004). This could especially be true for Ukraine which has experienced radical changes such as the dissolution of the Soviet Union and the ongoing land reforms. For that reason, historical drivers of land use change should be applied with care. In line with other land use change studies focussing on Europe, the agro-ecological suitability, the accessibility, the land conversion elasticity and the neighbourhood characteristics are included in the suitability factors for the allocation of cropland and pastures (Rounsevell et al. 2006; Verburg et al. 2006; Overmars et al. 2007; Verburg et al. 2008; Verburg and Overmars 2009; Britz et al. 2011). In addition to the agro-ecological suitability and the accessibility, Baumann et al. (2011) identified socio-economic factors that influence the post socialist land use changes in West Ukraine. The number of villages and the changes in population density and unemployment were correlated with agricultural land abandonment. It is assumed that these factors also apply for the rest of Ukraine and they have therefore been included in this study. In addition to these suitability factors, the current value of agricultural land expressed in the land rent prices (USAID 2006) is included as a proxy indicator for the areas where cropland and pastures will not be abandoned easily. It is assumed that agricultural land is more likely to occur in areas in the vicinity of other agricultural land (neighbourhood), markets (distance to city), and infrastructure (railroad, mostly used for agricultural commodities ); and in areas of high population density, little reduction in population, a high unemployment level and high land rental prices. Additionally, agricultural land is more likely to occur on land that is already in use for agricultural land (current land use), or on land that is easily converted to agricultural land use (elasticity) and has a high agricultural production potential (potential yield). The weighting of the suitability factors is difficult and expert consultation in the Ukraine is used to determine the relative importance of each suitability factor. Expert consultation is more often applied in land use change studies to weight the suitability factors, when longitudinal field observations are lacking (Overmars et al. 2007). Although, there was no consensus among the experts on the relative importance of each individual suitability factor, all experts indicated that agro-ecological suitability is the most important suitability factor. Table 6.2 depicts the suitability factors and the assigned weights used in the allocation procedure for agricultural land uses. The order of allocation of land use classes is: 1. Arable land 2. Mosaic cropland-pasture 3. Pasture land 4. Bioenergy crops or re-growth natural vegetation.

211

SHADES OF GREEN

No expansion of agricultural land (including energy crops) is allowed in the excluded areas. It is assumed that natural vegetation will re-grow on abandoned agricultural land in these areas. In the BAU scenario, agricultural land can expand in forest areas, in the progressive scenario all forest areas are protected and cannot be converted to agricultural land use. Autonomous deforestation is not included in the land use change modelling as deforestation is mainly related to illegal logging and clearing which is not recorded accurately (Kuemmerle et al. 2009). Moreover, this study focuses on changes in agricultural land use which is the dominant land use in Ukraine (75%), while forest occupies only 15% of the land area. The spatial data used to map land use changes are derived from different sources and are very heterogeneous with respect to level of detail, scale and quality. All datasets have been projected in Albers equal area projections in order to preserve the correctness of areas, the most important metric property in the model, as the model is build to assess 2 land use class areas. In addition, all maps have been resampled to a 1 km cell size. Data gaps have been filled with either the mean (for continuous data) or the majority (for categorical data) of a neighbourhood of 25 by 25 cells. Table 6.3 depicts the spatial data, the type of data and their source used in the land use allocation. Energy crop production For the modelling process it is assumed that the land that becomes available is taken into use for energy crop production in the same year that the land is abandoned. Abandoned agricultural land in the excluded areas (Chernobyl, conservation areas and slopes >16%) is assumed to be used for re-growth of natural vegetation. It is assumed that the energy crops are cultivated using state of the art cultivation practices. Therefore, high yields can be achieved directly from the beginning, and no learning (in terms of yield improvement over time) is assumed. The maximum attainable yield levels of wheat and switchgrass are derived from the Refuel study (de Wit and Faaij 2010; Fischer et al. 2010a; Fischer et al. 2010b). The spatial variation in yields due to the suitability is based on the agro-ecological suitability map. The cropland suitability map (FAO and IIASA 2007b) is used as a proxy for wheat, and the pasture suitability map (FAO and IIASA 2007a) is used as a proxy for the agro-ecological suitability of switchgrass cultivation.

212

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

Table 6.2: Suitability factors for land use allocation of cropland, mosaic cropland-pasture

and pasture.

Pasture

Explanation cropland-pasture

Land use

Cropland

Shape

Relationship

Direction

Suitability factors

Development at each individual location affects the conditions of neighbouring Neighbourhood + Linear 0.1 0.1 0.1 and distant locations. Some land use classes are more reluctant to change due to e.g. higher conversion Conversion elasticity 0.1 0.1 0.1 costs. The agro-ecological suitability includes climate soil and terrain constraints for Agro-ecological agricultural productivity. suitability + Exp. 0.3 0.25 0.2 Proxy indicator for the demand for land Land rent prices + Linear 0.2 0.2 0.2 in a specific region. The cattle density represents the + Linear 0 0.1 0.2 demand for grazing. Cattle density a The population density is an indication of Population density + Linear 0.05 0.05 0.05 the pressure on land. Due to decreasing population and Population change + Linear 0.05 0.038 0.025 urbanisation rural areas are abandoned. Unemployed people become more Unemployment + Linear 0.05 0.038 0.025 dependent on subsistence farming. Distance to cities represent the pressure Inv. from urban areas on their surroundings, Distance to cities + prop. 0.25 0.225 0.2 it is a proxy for distance to markets. The small plots in the surrounding of the village are most likely to remain in use compared to the plots further away from Distance to villages + Inv prop. 0.05 0.05 0.05 the village Accessibility is an important precondition for the use of agricultural land. The road Inv. density quite high and therefore no Distance to railroad + prop. 0.05 0.025 0 limiting factor for accessibility a Cattle is responsible for 96% of the total grass consumption (remainder is sheep en goat)(derived from FAO outlook), therefore the cattle density is assumed the best proxy-indicator for the spatial distribution of demand for grazing.

213

SHADES OF GREEN

Table 6.3: Spatial data input for PLUC model, the data source of the required datasets, their resolution and the adaptations made to these datasets. Suitability factors

Data source

Type of data

Land use

(ESA 2005)

Raster (300 m)

Suitability cropland

(FAO and IIASA 2007b)

Raster (5-arc minutes)

Suitability pasture

(FAO and IIASA 2007a)

Raster (5-arc minutes)

Suitability cropland-pasture

(FAO and IIASA 2007b; FAO and IIASA 2007a)

Raster (5-arc minutes)

Land rent prices

(USAID 2006)

Vector (Oblast level)a

Cattle density

(FAO 2005e)

Raster (0.05 decimal degree)

Population density

(FAO 2005a)

Raster (2.5 arc minute)

Population change

(State statistic service Ukraine 2011a)

Vector (Oblast level)a

Unemployment

(State statistic service Ukraine 2011b)

Vector (Oblast level)a

Cities

(Open Street Map 2010)

Vector

Railroad

(Pennsylvania State University Libraries 1997)

Vector

Inland water

(DIVA GIS 2011)

Vector

Chernobyl exclusion zone

(IUCN and UNEP-WCMC 2010)

Vector

Conservation areas

(IUCN and UNEP-WCMC 2010)

Vector

Excluded areas

Steep slopes (NASA and NGA 2000) Raster (30m) a Oblast is the NUTS 2 administrative level (equivalent to a province). Ukraine has 24 oblasts and one autonomous republic (Crimea).

Re-growth natural vegetation Without disturbance, natural vegetation can regenerate on abandoned agricultural land. The rate of regeneration and the maximum attainable biomass stock depends on the previous land use (the dense vegetation structure of grassland inhibits the establishment of new species including shrubs and trees (Benjamin et al. 2005; Verburg and Overmars 2009)), the suitability of the land for biomass production (Pueyo and Beguería 2007; Verburg and Overmars 2009), and the dispersal of seeds which is related to proximity of forest (Tasser et al. 2007; Verburg and Overmars 2009). This study only accounts for the suitability of tree generation, which is considered to be representative from a GHG and carbon stock perspective. The mean annual increment of forest was based on a map provided by Georg Kinderman of IIASA (personal communication). The amount of below ground biomass is calculated making use of root to shoot ratios of the IPCC (2006). The weighted average of the root to shoot ratios for young forest in a temperate climate based on the specie proportions of current standing stock was calculated, which in turn is derived from the country report of the global forest resource assessment (FAO 2010c). Although it is expected that it can take 40 to 100 years before forest reach full maturity (Houghton 1999; Kuemmerle et al. 2011), in this study it is assumed that the bulk of 214

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

carbon sequestration occurs in the first 20 years, as a linear growth rate applied in the model. The maximum total carbon stock achieved corresponds with 146 ton C /ha which is in line with the figures for mature forest (Houghton 1999; Kuemmerle et al. 2011).The total attainable biomass stock is assumed to be related to the spatial variation in the mean annual increment.

6.2.4 GHG emissions The net greenhouse gas balance is calculated on a grid cell level, taking into account CO2 emissions from carbon stock changes, N2O and CH4 emissions, and GHG emission abatement by means of substituting fossil fuels for biofuels. All four variables are spatial and temporal, meaning that they vary over x, y and t. Equation 6.1 depicts the overall GHG balance.

44    44  GHGbalance =  ∆C ⋅ − ⋅ GWPCO2  +  N ⋅ ⋅ GWPN 2O  + ( CH4 ⋅ GWPCH 4 ) − GHGabatement 12 28     Equation 6.1 GHG ∆C GWPCO2 NN2o GWPN2O CH4 GWPCH4 GHGabatement

Net Green house gas emissions Change in carbon stock Global warming potential CO2 N emitted in the form of N2O Global warming potential N2O Methane emissions Global warming potential CH4 Avoided GHG emissions by replacing fossil fuels

Kg CO2-eq/ha/yr Kg/ha/yr factor Kg/ha/yr factor Kg/ha/yr factor Kg/ha/yr

The global warming potentials (GWPs) are derived from (IPCC 2007c) and assume a time horizon of 100 year. Changes in carbon stock When land is converted from one land use to another or when land use management changes, carbon can accumulate (carbon sequestration) or diminish (carbon emissions). In this study, the IPCC approach (IPCC 2006) to calculate CO2emisisons related to changes in carbon stocks is applied which involves five carbon pools: above-ground biomass, belowground biomass, dead wood, litter, and soil organic matter. Since the Tier 1 approach of the IPCC assumes an equilibrium in the carbon stocks in dead wood and litter in cropland and pasture, these carbon pools are not considered in this study. The organic carbon content of soils is related to the soil type, climate, land use and applied agricultural management and is therefore spatially highly heterogeneous. Land 215

SHADES OF GREEN

use and management changes affect the soil organic carbon content: e.g. a permanent vegetation cover reduces respiration of the soil and therefore the oxidation of soil organic carbon, whereas intensive tillage and drainage increases the loss of soil organic carbon. The application of organic inputs such as manure and crop residues could increase soil organic carbon (SOC). depicts the relations between the soil organic carbon content and other carbon pools.

Figure 6.1: Schematic overview of carbon exchange between atmosphere, biomass and soil.

For every year, the spatially explicit change in SOC is calculated according to Figure 6.2:

 SOCt − SOCt −1  = ∆C SOC  + ∆COrganic  D  Mineral Equation 6.2 ∆CSOC SOCt SOCt-1 D ∆Corganic

Annual change in SOC SOC at time step t SOC at previous time step Time required for equilibrium Change in SOC in organic soil

Ton C/ha/year Ton C/ha Ton C/ha Years Ton C/ ha/year

The change in carbon in organic soils is a fixed annual carbon flux (in ton/ha/yr) depending on the land use and climate region. The SOC in mineral soils at time step t is calculated given the soil type, climate, land use, management and previous conversions within the modelling timeframe based on default values of the IPCC (2006). The climate region and soil type are spatially explicit and static. The spatial attribution of the climate regions is based on the IPCC climate classification related to average annual precipitation and temperature (IPCC 2006). Maps of annual averages of precipitation and temperature were 216

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine 13

constructed by interpolating (Renka 1987) the precipitation data of 45 domestic and 5 neighbouring weather stations derived from (NOAA 2008). Spatial explicit data on soil characteristics is derived from (Batjes 2005) and reclassified to the IPCC soil classes making use of the reclassification method proposed by IPCC (2006). The spatially explicit land use for year y is the key output of the PLUC model (section 2.3). The management level of the agricultural land use including the tillage regime and the organic carbon application levels are scenario specific. The land use, management and organic input factors affecting the SOC content are depicted in Table 6.4 of the Appendix (Section 0). Some studies indicate that it could take up to 50-100 years to reach a new SOC equilibrium (Kuikman et al. 2005). However, in this study a time horizon of 20 years is assumed in line with the IPCC (2006), and as proposed by the EC (2008) and NTA 8080 (NEN 2009). The carbon stocks in above and below ground biomass depend on the land use, the productivity of the land and the management applied. For annual crops, the increase in biomass stocks in a single year is assumed to be equal to biomass losses due to harvest and degeneration in that same yea: there is no net accumulation of biomass carbon stocks in annual arable crops (IPCC 2006). For pastures and switchgrass, the biomass carbon stock is equal to the carbon content of the below ground biomass (it is assumed the above ground biomass is removed due grazing or harvested for consumption). As <3% of the total arable land of Ukraine is in use for permanent crops (FAO 2010a), the carbon accumulation in permanent crops is not considered in this study. For natural vegetation such as forest, shrubland and natural grassland the carbon in both the above and the below ground biomass are included. The C stock in biomass changes at once, when one land use is converted to another. A SOC equilibrium time of 20 years is assumed, which implies a continued SOC change after land use conversions. Nitrous oxide emissions Fifty to sixty percent of the anthropogenic induced N2O emissions originates from agriculture, the largest share being direct emissions from agricultural soils (Mosier et al. 1998). In this study, the nitrous oxide emissions from agricultural soils and from the livestock sector are included in the overall modelling approach. The N2O that is formed during the nitrification and denitrification processes in the soil is emitted to the atmosphere. The amount of nitrous oxide emitted is related to the amount of inorganic of mineral nitrogen available. The IPCC guidelines (2006), propose a default emission factor (EF) of 1% for nitrogen inputs from mineral fertilisers, organic amendment, and crop 13

The tension spline interpolation method is used in order to preserve e.g. positivity and monotonicity (Renka, 1987). 217

SHADES OF GREEN

residues. Many studies regarding GHG emission from energy crop production or GHG in agriculture in general, apply this default emission factor (Smeets et al. 2009a; de Wit et al. 2011c; Popp et al. 2011). However, several studies have indicated that N2O emissions vary to a great extent depending on local conditions such as agricultural land use, soil type and N-origin (Stehfest and Bouwman 2006; Britz and Leip 2009; Smeets et al. 2009a; Lesschen et al. 2011). Therefore, this study models the N2O emissions in detail spatially explicitly. N2O emissions from agricultural soils are directly related to the amount of mineral nitrogen available, the source of mineral nitrogen and the biophysical conditions. In this study, balanced fertilisation is assumed, in line with the MITERRA-EUROPE model (Velthof et al. 2009). Balanced fertilisation implies fertiliser and manure application rates in accordance with the nitrogen crop demand after accounting for the crop uptake factor, nitrogen inputs from grazing, atmospheric deposition, mineralisation, biological N fixation and accounting for the losses due to direct N2O emissions, volatisation (NH3 and NOx), runoff and leaching. The losses of nitrogen trough direct N2O emissions, volatisation and run-off are nitrogen source and condition specific. The indirect N2O emissions are related to the amount of nitrogen that is lost due to run-off, leaching and volatisation. For that reason, the total N application and therefore the N2O emissions can only be calculated when all nitrogen losses trough other pathways are included in the equation. Also N2O emissions from housing and storage of manure are included. As no point sources of GHG emissions are taken into account, all housing and storage of manure are allocated to the manure which is distributed over agricultural land in proportion to the nitrogen-gap between nitrogen requirements and nitrogen supplied by other sources. Figure 6.2 provides an overview of the N pathways addressed in this study. Equation 6.3 (adapted from Velthof et al. (2009)) provides the calculation method for balanced fertilisation and indicates the relation between the nitrogen sources and nitrogen losses.

(N

crop

(N

(

)

+ Nres ) ⋅ fuptake ,crop−res = N fert − NN2O , fert − Nvol , fert − Nrunoff , fert +

m−man

)

(

− NN2O ,man − Nvol ,man − Nrunoff ,man + fgraz ⋅ Nm−graz − NN2O ,graz − Nvol ,graz − Nrunoff ,graz

((

) (

) (

)

+N fix + fdep ⋅ ( Ndep − NN 2O ,dep ) + fmin ⋅ NRres − NN2O ,Rres + No−man − NN2O ,min + NSOM − NN2O ,min

))

Equation 6.3 Ncrop Nres fuptake, crop-res Nfert NN2O, fert Nvol,fert Nrunoff,fert

N in crop N in residue (above and below ground) Crop uptake factor N supplied by synthetic fertilizer N in N2O emissions from synthetic fertilizer N volatised from synthetic fertilizer N in runoff from synthetic fertilizer 218

Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha

6.

Nm-man NN2O, man Nvol,man Nrunoff,man fgraz Nm-graz NN2O,graz Nvol,graz Nrunoff,graz Nfix fdep Ndep NN2O,dep fmin NRres NN2O,Rres No-man NN2O,min NSOM

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine Mineral N supplied by manure N in N2O emissions from mineral N in manure N volatised from mineral N in manure N in runoff from mineral N in manure Available fraction of grazing deposits Mineral N in grazing deposits N in N2O emissions from mineral N in grazing deposits N volatised from mineral N in grazing deposits N in runoff from mineral N in grazing deposits Biological N fixation Available fraction of atmospheric deposition N supplied by atmospheric deposition N in N2O emissions from atmospheric deposition Available fraction of mineralised N Organic N supplied from residues left on the field (above and below ground) N in N2O emissions from residues Organic N supplied by manure N in N2O emissions from mineralisation of organic N Mineralised N related to soil organic carbon loss

Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha Kg/ha

Figure 6.2: Schematic overview of N pathways modelled in this study. Adapted from Velthof et al (2009).

In this study, the method to assess the nitrogen pathways from agricultural soils described by Velthof et al. (2009), is applied. However, some alterations to the MITERRA method (Velthof et al. 2009) have been made: land use, soil type and nitrogen-source specific N2O 219

SHADES OF GREEN

emission factors are applied (and included in the equation for balanced fertilization, instead of the default emission factor) and a the coupling between the SOC losses and nitrogen-mineralisation is made. Moreover, in this study, the nitrogen related emissions are modelled both spatially explicitly (on a detailed level, allowing for differentiation in emissions for spatial heterogeneity in biophysical conditions) and dynamically (to allow for modelling of nitrogen emission dynamics over time). Hence, all variables in Equation 6.3 vary over space and time. The total N2O emissions are the sum of direct N2O emissions from the nitrogen supplied by fertilizer, manure, grazing, atmospheric deposition, residues and mineralisation; and the indirect N2O emissions from the volatisation, runoff and leaching from the N from fertilizer, manure and grazing (see Equation 6.4). NN2O =

(N

N2O , fert

)

+ NN2O ,m−man + NN2O ,m−graz + NN2O ,dep + NN2O ,Rres + NN2O ,min +

fN2O ,vol ⋅ ( Nvol , fert + Nvol ,man + Nvol ,graz ) + fN2O ,runoff −leach ⋅ ( Nrunoff , fert + Nrunoff ,man + Nrunoff ,graz + Nleach ) Equation 6.4

NN2O fN2O,vol fN2O,runoff, leach Nleach

Total N2O emissions Fraction of volatised N that results in N2O emissions Fraction of leached and runoff N that results in N2O emissions N that leaches below the rooting zone

Kg/ha Kg/ha

More detailed information on the method of analysis of the individual N-pathways and the data used in this assessment are described in van der Hilst et al. (2012b). The direct and indirect N2O emissions factors differentiated for land use, soil type and nitrogen source, used in this study are derived from Lesschen et al. (2011) and IPCC (2006) and are depicted in Table 6.5 of the Appendix (Section 6.6.1). Methane emissions Methane emissions from enteric fermentation and manure storage depend on the number and type of livestock, their enteric fermentation rate, and the type of storage facility. The number of animals per specie is derived from (FAO 2011; State Statistics Service of Ukraine 2011). In this study the country specific data on methane emissions from enteric fermentation and from housing and storage per head are derived from the Ukraine country report of the UNFCC (2011). As no point sources of GHG emissions are taken into account, all methane emissions are allocated to the manure which is distributed over agricultural land in proportion to the N-gap between N requirements and N supplied by other sources.

220

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

Avoided emissions due to replacement of fossil fuels The GHG emission reductions resulting from the replacement of fossil fuel are allocated to the areas of energy crop production. The total abatement is calculated taking into account the cumulative biomass yield, the conversion efficiency and the abatement figures per litre of ethanol which are derived from JRC (JEC 2008). The spatial allocation of the GHG abatement is directly related to the crop yield and the conversion efficiency from the crop to ethanol. The cultivation related N2O emissions have been excluded from the abatement figures as the N2O emissions are already accounted for. Table 6.6 of the Appendix (Section 6.6.1) depicts the abatement figures for ethanol from wheat and switchgrass. Mitigation measures of GHG emissions from agriculture Measures to reduce GHG emissions are reduced tillage, increased carbon input trough less residue removal and fertilizer type improvement. Reduced tillage reduces both carbon and nitrous oxide emissions. Increasing the amount of residues on the field increases soil organic carbon and reduces the required nitrogen inputs from manure and fertilizers. However, the direct nitrous oxide emissions from residues will increase. Shifting from a mixture of nitrate based and ammonium based fertilizers to ammonium based fertilizer only decreases the direct nitrous oxide emission from fertilizer application. The mitigation measures are based on the measures proposed in Lesschen et al. (2008), de Wit et al. (2011c) and Smith et al. (2000; 2008). The land use, management and organic input factors affecting the SOC content are depicted in Table 6.4 of the Appendix (Section 0). The factors are differentiated for the BAU and progressives scenario and for the progressive scenario taking mitigation measures. The nitrous oxide emission factors of residues and ammonium based and nitrate based fertilizers are depicted in Table 6.5 of the Appendix (Section 6.6.1).

6.3 Results 6.3.1 Total agricultural land balance In Figure 6.3, the developments in production of food and feed in million ton dry weight is depicted up to 2030 for the BAU and progressive scenario. Although it is assumed that the increase in consumption is the same in the BAU and in the progressive scenario, the demand for feed is lower in the progressive scenario as the livestock sector becomes more efficient. Therefore, less feed input is required for the same meat and milk output in the progressive scenario. In addition, in the progressive scenario a shift towards more feed crop consumption at the expense of grass consumption is assumed. Therefore, the total crop demand is higher and the total grass demand is lower in the progressive scenario 221

SHADES OF GREEN

compared to the BAU scenario. However, differences between the two scenarios in terms of total production are limited.

Figure 6.3: Development in demand for domestic produced food and feed in the BAU and progressive scenario in million ton dry weight product.

In Figure 6.4, the developments in crop and pasture yields and the efficiency in the livestock sector are presented for the two scenarios compared to the levels of 2010. It is clear that the productivity increase is close to zero in the BAU scenario, whereas in the progressive scenario the productivity increases rapidly; especially the crop and pasture yields. Figure 6.5 presents the total land requirements for crop production and grazing given the demand depicted in Figure 6.3 and the productivity presented in Figure 6.4 and assuming an average agro-ecological suitability of cropland and pasture equal to the average suitability of the cropland and pasture currently in use. The currently low cropping intensity indicates that considerable land areas of cropland and pasture are left fallow. In the progressive scenario, land is used more efficiently and no land is left fallow by 2030. Because of the higher yields and the higher cropping intensity, only half of the land currently in use for agricultural production is required to meet the demand in 2030. However, the actual land area depends on the location-specific productivity of the land (the agro-ecological suitability) and therefore of the location of production. For that reason, developments in actual land requirements and land availability for other land use functions can only be assessed using a spatiotemporal land use change model.

222

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

Figure 6.4: Developments in crop and pasture yield and livestock productivity for BAU and progressive scenario compared to 2010 levels (2010= 1).

Figure 6.5: Developments in land requirements for crop production and grazing for BAU and progressive scenario based on the food and feed requirements, the yield and efficiency development of the two scenarios and assuming the current average agro-ecological suitability of arable land and pastures.

6.3.2 Spatially explicit results Figure 6.6 depicts spatially explicitly the land use for the BAU and progressive scenario for 2010, 2020 and 2030. In both scenarios, the growth rate of the land area for crops exceeds the growth rate of pastures and there is a tendency towards more dedicated cropland and pasture land, at the expense of mosaic cropland-pasture. In the BAU scenario, cropland 223

SHADES OF GREEN

expands at the expense of mosaic cropland-pasture in areas with a high agro-ecological suitability and which are currently already popular agricultural areas (have high land rent). This is a clearly visible in the central west region (Ternopil, Vinnytsia, Cherkasy and Kiev oblast). Expansion of agricultural land is mostly at the expense of grassland and shrubland, and sometimes at the expense of forest especially the patches that are surrounded by agricultural land and are located in very suitable areas. Non-agricultural land is often first converted to pasture and subsequently to cropland-pasture and cropland. In the BAU scenario little land is available for bioenergy crops as all current agricultural land is required to meet the future food and feed production demand. In the progressive scenario, agricultural land is rapidly abandoned due to the high productivity increase. Land is primarily abandoned in the areas which are less suitable for agricultural production such as the north and north-western areas, which are mainly marshy mosaic forest areas, the south west, which are the Carpathian Mountains (cropland), the eastern areas which are more industrialized, and the south, which receives little precipitation. Agricultural production concentrates in the central parts of Ukraine which are most suitable for both crops and pastures. In addition, these areas are currently popular for agricultural production, have relative high population density, are in the vicinity of large cities, have many villages, have access to railroads and have high unemployment rates. In the progressive scenario there is a trend towards more dedicated cropland and pasture at the cost of mosaic cropland-pasture. Therefore, cropland-pasture areas are abandoned more rapidly than pure cropland, and dedicated pastures areas even expand slightly. Although the conversion elasticity is in favour of converting abandoned agricultural land to pasture, pasture areas are expanding in other areas which are more favourable in terms of agro-ecological suitability, at the expense of shrubland. In Figure 6.7 the developments in land use requirements for the BAU and progressive scenario is depicted for 2010, 2020 and 2030 as they result from the spatially explicit modelling. Figure 6.5 is constructed based on the food and feed requirements, the agricultural efficiency of the two scenarios and taking into account the average agroecological suitability of the land currently in use of cropland and pasture. Figure 6.7 is constructed based on the spatial modelling using the same food and feed requirements and scenario assumptions as Figure 6.5 is based on, but taking into account the spatial variability of the agro-ecological suitability of cropland and pasture. For the BAU scenario, it is clear that the modelled land requirements are higher compared to the level depicted in Figure 6.5. This is due of the fact that the agricultural land is forced to expand to areas that are less suitable for agricultural production (as all favourable areas are already in use). For the progressive scenario, the modelled land requirements to meet food and feed productions are significantly lower compared to the levels depicted in Figure 6.5. 224

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

Therefore, more land will become available for energy crop production (32.1 Mha). This is a logical consequence when considering that agricultural land use concentrates in the most suitable areas. These results illustrate the importance of assessing land use dynamics and related land availability for energy crops spatially explicitly. Figure 6.8 shows the development in potential annual biomass feedstock production of switchgrass (whole plant) and wheat (grain only) for the period up to 2030. Although the assumed conversion efficiency from wheat to ethanol is higher than from switchgrass to ethanol, the potential ethanol yield per hectare is higher for switchgrass due to the higher biomass yields (a maximum yield of 170 GJ/ha/yr for switchgrass and 100 GJ/ha/yr for wheat). In the progressive scenario, up to 5.0 EJ biomass could be produced on the available land (2030) compared to the potential wheat production 3.6 EJ (grain). As in the BAU scenario little land becomes available, potential annual production is low compared to the progressive scenario (±2 PJ for wheat and switchgrass in 2030).

225

SHADES OF GREEN

Figure 6.6: Land use for 2010-2020-2030 for the BAU (left) and progressive scenario (right).

226

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

Figure 6.7: Developments in land requirements for crop production and grazing for the BAU and progressive scenario based on the spatially explicit modelling of the land use change in the period 2010-2030 using the PLUC model.

Figure 6.8: Development in potential annual biomass production (whole crop for switchgrass and grain only for wheat) for the BAU and progressive scenario in PJ/yr.

6.3.3 GHG emissions Developments in GHG emissions are calculated spatially explicitly on an annual basis on a 2 spatial resolution of 1 km . This results in annual maps for the timeframe 2010-2030 of N2O, CO2 and CH4 emissions, as well as the avoided emissions related to the replacement of fossil fuels. The emissions are calculated for the BAU scenario, for the progressive 227

SHADES OF GREEN

scenario, and for the progressive scenario taking into account the management measures to reduce CO2 and N2O emissions. In the Appendix (Section 6.6.2), snapshots of annual nitrous oxide emissions, carbon stock changes, and CO2 abatement for the years 2010, 2020 and 2030 are presented for the progressive scenario (due to the fewer land use changes, the changes in GHG emissions in the BAU scenario are less significant, and therefore not shown). Annual average GHG emissions and sequestration In Figure 6.9 the average annual GHG emissions (including N2O, CO2 and CH4 and the avoided GHG emissions) for the timeframe 2010‐2030 relative to the levels in 2009, are presented spatially explicitly for the progressive scenario. What is most apparent fromFigure 6.9, is that in some areas GHG emissions increase, whereas in other areas GHG emissions decrease. Increased emissions occur in all areas that remain in use for agricultural production for food and feed. As production intensifies, fertiliser use per hectare is increased, which results in higher direct and indirect nitrous oxide emissions. In addition, it is assumed that more intensive agricultural management also implies full tillage resulting in significant carbon losses on arable land. It is assumed that in the progressive scenario the pastures are managed better compared to the levels of 2009, resulting in higher yields (and therefore larger carbon stocks in below ground biomass) and higher soil organic carbon content but also on higher N2O emissions. Large GHG emission reductions occur when abandoned cropland is used for switchgrass or regeneration of natural vegetation. The outliers of high GHG emissions (i.e. >1 tonne CO2-eq per hectare per year, representing ±1% of the area) occur in a few areas in the southern, the north western (in case of wheat), and central part of Ukraine. The outliers in the south are a result of the conversion of shrubland to pastures and to mosaic cropland-pasture, which results in soil organic carbon losses, aboveground biomass loss (depending on the local yield of shrubland and pastures) and an increase in N2O emissions due to grazing depositions (and manure and fertilizers when nitrogen requirements exceed the grazing depositions). The outliers in central Ukraine are the result of the shift from mosaic cropland-pasture to cropland which results in a reduction in below ground biomass and soil organic carbon and an increase in direct and indirect N2O emissions from increased nitrogen requirements from cropland compared to pastures. The high increases in GHG emissions in the north western area only occur when wheat is cultivated on the abandoned land. The high GHG emissions are caused by the significant increase in annual carbon fluxes and the related N2O emissions of the organic soils that are present only in that part of Ukraine when abandoned pastures are converted to wheat. High sequestration potentials (i.e. >10 tonnes CO2-eq /ha/yr) are found in 1% of the areas when the abandoned land is used for 228

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

re-growth of natural vegetation, 0.1 % when the land is used for wheat (the outliers are in the excluded areas where no wheat is cultivated but natural vegetation re-grows) and in 3.5% of the areas when switchgrass is cultivated on the abandoned agricultural land. The high sequestration potentials are mainly achieved in high productivity areas on clayey and organic soils. Re-growth natural vegetation When the land is used for the re-growth of natural vegetation, a considerable amount of carbon is stored in above and below ground biomass, as well as through an increase in soil organic carbon. The highest carbon accumulation levels are reached in the areas with the highest mean annual increment of forest and in moist areas with clayey soils. It should be noted that not all soil organic carbon sequestration is accounted for, as an equilibrium time of 20 years is assumed, and some areas will only become available at the end of the timeframe assessed. Therefore, the carbon sequestration will continue for several years after 2030, but will not go on indefinitely, as equilibrium of soil organic carbon will be reached, and as mature forests will not sequester biomass carbon anymore. How long it will take until this equilibrium in forest is reached depends on several local biophysical factors, the tree species and the forest management. 90% of the area has an average annual GHG emission or sequestration between -8.8 and +0.7 with an overall average of 2.9 ton CO2-eq per hectare per year. The total net average annual GHG balance for the period 2010-2030 is -172 Mton CO2-equivalent. Wheat Although a net negative GHG balance is achieved when wheat for bioethanol is cultivated on the abandoned land, fewer GHG emissions are avoided in the timeframe 2010-2030 when compared to the use of abandoned land for re-growth of natural vegetation or switchgrass. The N2O emissions are significantly higher, as in addition to the food and feed crops, wheat for bioethanol requires nitrogen input. Therefore, the total nitrogen input and the related direct and indirect N2O emissions are much higher compared to re-growth of natural vegetation. In addition, pastures are converted to wheat, which results in carbon losses due to a loss of soil organic carbon and below ground biomass. However, as it is assumed that currently the pastures are not well managed and are therefore semidegraded, and that wheat is cultivated leaving 50% of the residues, the losses in organic carbon are limited. In addition, significant parts of the abandoned agricultural land are located in the ‘excluded areas’, which are assumed unsuitable for energy crop cultivation. The re-generation of natural vegetation in these areas results in a considerable carbon stock in above and below ground biomass and SOC, which compensates the losses of biomass and SOC in the areas where pasture is converted to wheat. 90% of the area has an average annual GHG emission or sequestration ranging between -2.7 and +0.4, with an

229

SHADES OF GREEN

average of -0.7 ton CO2-eq per hectare per year. The total net average annual GHG balance is -45 Mton CO2-equivalent. Switchgrass The use of abandoned land for the cultivation of switchgrass for bioethanol production results in the most favourable GHG balance. Nevertheless, the N2O emissions are the highest for this variant compared to the use of abandoned land for re-growth of natural vegetation or wheat. This is the result of the high yields of switchgrass and therefore the high nitrogen requirements (despite of relative low nitrogen content of the crop). However, the conversion of cropland to switchgrass results in high increase in soil organic carbon (due to the no tillage of switchgrass), and increase in below ground biomass. Also the conversion from the currently poor managed pastures to well managed switchgrass results in an increase in soil organic carbon, and due to the higher yields compared to pastures also in an increase in below ground biomass. Moreover, the higher yields and the higher abatement value result in higher avoidance of GHG emission by the replacement of fossil fuel. When the abandoned land is used for switchgrass, 90% of the average annual GHG emission or sequestration varies between -9.6 and 0.4 with an average of -3.1 ton CO2-eq per hectare per year. The net average annual GHG balance is -191 Mton CO2equivalent. It should be noted that the gains and losses in organic carbon will reach an equilibrium, but that GHG abatement can continue. The annual N2O, CO2 emissions and GHG-abatement is spatially depicted for the years 2010, 2020 and 2030 for the progressive scenario are in Figure 1.11 in the Appendix (Section 1.6.2). Cumulative GHG emissions and sequestration The developments in GHG emissions are assessed spatially explicitly on an annual basis for the timeframe 2010-2030. The results of these assessments are summarised in graphs of the total cumulative emissions for the timeframe 2010-2030 in Figure 6.10 for the BAU scenario (1), the progressive scenario (2) and for the progressive scenario + abatement measures (3). The three variants with the use of the abandoned agricultural land for regrowth of natural vegetation (a) and the use for wheat (b) and switchgrass (c) for bioethanol production are considered for all scenarios. In the BAU scenario there are small differences between the three alternative usages of the abandoned agricultural land, as little land becomes available. The carbon emissions increase due to the expansion of agricultural land at the expense of shrubland and forest and the conversion of mosaic cropland-pasture to dedicated cropland. The carbon emissions are mainly from the loss in above and below ground biomass (85%); soil organic carbon losses play a minor role. The N2O emission increase slightly due to higher overall production and related nitrogen requirements. The total cumulative emissions for the timeframe 2010-2030 are ± 1.1 GT CO2-equivalent. 230

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

The progressive scenario differs considerably from the BAU scenario, and also the three alternative land usages result in significant differences in cumulative emissions. The most apparent developments are already explained in the description of Figure 6.9. The use of abandoned land for re-growth of natural vegetation results total cumulative GHG emissions avoidance of 3.5 GT CO2-equivalent in the timeframe 2010-2030. When the abandoned land is used for wheat for bioethanol production a total of 0.8 GT CO2equivalent can be avoided. For switchgrass the GHG emission reduction amounts 3.8 GT CO2-equivalent. When measures are taken to reduce soil organic carbon losses and N2O emissions, higher net GHG emission reductions can be realised. As it is assumed that reduced tillage is applied on arable land and that pastures are improved, soil organic carbon levels increase on all agricultural land resulting in 9 % (for wheat) to 44% (for switchgrass) more cumulative carbon sequestration compared to the regular progressive scenario. In addition, it is assumed that only ammonium based fertilisers are used (instead of a mixture of ammonium and nitrate based fertilisers), which results in lower direct N2O emissions from synthetic fertilisers. The cumulative N2O emission reductions vary between 3% (for re-growth natural vegetation) and 14% (for switchgrass) compared to the levels of the regular progressive scenario. The total cumulative GHG balances for the timeframe 2010-2030 are -2.6 GT CO2-equivalent for wheat cultivation, -4.3 GT CO2equivalent for re-growth of natural vegetation, and -5.0 GT CO2-equivalent for switchgrass. For all scenarios the total cumulative methane emissions are equal, since for all scenarios the same number of animals, the same enteric fermentation per animal and the same manure storage facilities are assumed. The spatial results show differences in the spatial distribution of methane emissions as these emissions are allocated to the manure and manure is spread according to the gap between the local nitrogen requirements and the local nitrogen supplied from other sources.

231

SHADES OF GREEN

Figure 6.9: Average annual GHG balance of timeframe 2010-2030 compared to levels of 2009 for the progressive scenario when abandoned agricultural land is used for re-growth natural vegetation 2 (top), wheat for bioethanol (middle), or switchgrass for bioethanol (bottom) in ton CO2-eq/km per year. 232

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

Figure 6.10: Graphs of development of cumulative GHG emissions for the timeframe 2010-2030 for the BAU scenario (1), the progressive scenario (2) and the progressive scenario with abatement measures (3). The three variants considered are the use of abandoned land for re-growth of natural vegetation (a), the use for wheat for bioethanol (b) and the use for switchgrass for bioethanol (c).

6.4

Discussion

6.4.1 PLUC model Ukraine is a country in transition: how, in which direction, and in what pace Ukraine will develop is highly uncertain. This makes assessments of future land use developments relatively difficult as historical drivers of change may no longer apply and new drivers are likely to arise. Both the developments in the drivers of national production (such as population, consumption per capita, imports and exports) and the drivers of efficiency improvement (technology adoption) are uncertain. In addition, the suitability factors for land use change (such as population density, conversion elasticity, agro-ecological suitability) and their relative importance are uncertain and could be counteractive. PLUC can partly deal with these uncertainties by enabling Monte Carlo analysis on stochastic

233

SHADES OF GREEN

input parameters as is described by Verstegen et al. (2011) and van der Hilst et al. (2011). However, as the aim of this study is to develop and provide the proof of principle of the spatiotemporal dynamic GHG emission module of PLUC, the LUC has only been assessed for two highly divergent scenarios, acknowledging the high uncertainties. The scenarios selected could be perceived as quite extreme: almost no improvement in productivity and a slight increase in agricultural area in the BAU scenario and a very high increase in productivity and a related strong reduction of required agricultural land in the progressive scenario. The most likely developments will be somewhere in the middle. Wheat and switchgrass are selected as bioenergy crops in to compare the GHG emission abatement potentials for a typical first and second generation ethanol crop. However, as Ukraine is perceived to be the breadbasket for Europe and even on a global scale, there could be a psychological barrier for Ukrainians to produce wheat for energy purposes. As bioenergy crop production was not included as a dynamic land use class, the competition for land or indirect land use change as a result of bioenergy crop production has not been modelled in this study. As the starting point of this study is that competition for land is to be avoided, land in use for food and feed production was excluded from conversion to energy crop production. It is assumed that the land that becomes available by means of more efficient production could be used for bioenergy production. Naturally, the land could also be used for other purposes such as increased food and feed production for increased exports. However, the aim of this study was to assess the potential and GHG impacts of the implementation of bioenergy production without threatening the feed and food production. It is assumed that the agricultural practices and the adoption of improved practices and related increases in efficiency are uniform for the whole country as well as uniform for all different kind of producers. However, there are currently large differences in agricultural practices, the commodities typically produced, and distribution of small and large scale producers. It is expected that they will not all develop in the same pace. However, for the sake of the land use change modelling on national scale, an average agricultural practice and an average increase in efficiency is assumed for the entire country.

6.4.2 GHG calculations Carbon stocks The changes in soil organic carbon make a major contribution to the total GHG balance. Spatial information available on soil organic carbon content have a coarse resolution, are old and are not linked with current land use data. As the land use has a high influence on 234

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

soil organic carbon, the independent use of these datasets could cause significant errors. For that reason it is assumed that the IPCC method is the best approach at this point. This is also applied in other (non-dynamic) models on GHG emissions from agricultural land use, such as the MITERRA model (Lesschen 2008). The IPCC provides reference SOC values and emission factors related to land use, management, and inputs related to the climate region. The choice for specific management and input factors are key for calculated soil organic carbon content. However, these factors are rather rudimentary, only linked to a descriptive definition of management and input level, and are not specific for Ukraine. Historic land use and management has a significant effect on SOC. However, due to a lack of data, historic land use prior to the modelling period has not been taken into account. Although some studies indicate that it could take up to 50-100 years to reach a new SOC equilibrium (e.g. Kuikman et al. 2005), a time horizon of 20 years is assumed in line with the IPCC. This assumed required time to reach equilibrium is a key factor for the order of magnitude of the carbon emission/sequestrations resulting from changes in SOC. As the soil organic carbon has such a large contribution to the overall GHG balance, it would be preferable to use up to date field data of soil organic carbon stocks and changes related to land and management changes. The changes in carbon stock in biomass also have a significant contribution to the total GHG balance. In this study, the default value for root to shoot ratio of 4 has been applied for grassland and switchgrass. However, the root to shoot ratio can depend on local agroecological conditions. Therefore, the root to shoot ratio of switchgrass and the related carbon sequestration in below ground biomass could be over or underestimated. In this study, the below ground biomass of switchgrass is accounted for as a carbon stock similar to grassland. It is however expected that after the lifetime of switchgrass, roots will decay. This would imply a loss of carbon in line with the decay of below ground residues of annual arable crops for which no biomass carbon stock is accounted. On the other hand, if the land would be replanted with switchgrass, the carbon stock of below ground biomass would be re-established. As the same applies for the renewal of grassland, the carbon stock in below ground biomass of switchgrass is included in line with the assumptions for grassland. The mean annual increment of above ground biomass in forest is based on a map on net primary production provided by IIASA. This could result in an under estimation of the biomass growth rate of regenerated forest as the data is based on the accumulation of standing stock, while it is assumed that young forest will grow faster than mature forest. In addition, the data does not take the biomass from leaves and branches <7cm into account. On the other hand, re-growth of forest could be suppressed due to the slow penetration of the previous vegetation (pasture). As the carbon stock play an important 235

SHADES OF GREEN

role in the overall GHG balance, more detailed spatially explicit data on growth rates of above and below ground biomass is required. Currently, large parts of straw residues are burned on the field in Ukraine. It is however not recorded how much and where straw is burned. For that reason it was not included in this study, but it would have a negative impact on the GHG balance. The GHG balances are calculated for the timeframe 2010-2030. However, when a longer timeframe would be considered, the GHG balances would look quite different. If it is assumed that no more land use or management changes will occur after 2030, the carbon stock related emissions are phased out over time as a result of the assumed SOC equilibrium and the fact that mature forest will no longer sequester carbon. Assuming an equilibrium time of 20 years for both SOC and re-growth natural vegetation, in the progressive scenario the peak in carbon sequestration in soil and biomass is around 2030 and to levels out towards 2050. Thereafter, it is expected that no additional carbon is sequestered in soil or biomass. However, avoidance of carbon emissions due to the replacement of fossil fuels will continue over the years. In addition, the N2O and methane emissions will also continue. Considering a total timeframe to 2100, the total cumulative GHG mitigation is much higher for all abandoned land use alternatives. When the abandoned land is used for re-growth of natural vegetation, the GHG emission reduction could increase from -3.5 in 2030 to ± -5.5 Gt CO2-eq in 2100 in the progressive scenario. However, the use of the abandoned land for energy crops result in much higher GHG emission reductions: wheat could achieve ± -6 Gt CO2-eq by 2100 compared to the -0.8 Gt CO2-eq in 2030. Cultivating switchgrass on the abandoned land could even result in a GHG emission reduction of -15 Gt CO2-eq by 2100 which is almost three times higher than when the land is used for the re-growth of natural vegetation. These emissions estimates include all GHG emissions of all agriculture land and livestock, and include re-growth of natural vegetation in areas where agricultural land is abandoned but are excluded for energy crop production. N2O emissions The N2O emissions in Ukraine that are calculated in this study are relatively low compared to agricultural nitrous oxide emissions in other European countries found other studies (Leip et al. 2008; Velthof et al. 2009; de Wit et al. 2011c). This has several reasons: In this study it is assumed that balanced fertilization is and will be applied. It is however likely that currently both under and over fertilization is applied and that therefore emissions are higher at some locations and lower at others than modelled in this study. In addition, it could be expected that in a progressive

236

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

scenario, over fertilization will be applied like it has been /is done in other parts of Europe. Therefore the N2O emissions could be under estimated. In several studies the default N2O emission factor of 1% of N-application is used. In this study the soil, land use and N-source specific emission factors derived from Lesschen et al. (2011) are applied. These emission factors are almost all lower than the 1% default value. The N2O emissions of grazing and (stored) liquid manure are high. However, the Ukrainian livestock sector is relatively small compared to other European countries and most of the manure is stored in solid systems. In line with the IPCC (2006) it is assumed that nitrogen is mineralised when SOC is diminished, but no organic nitrogen is accumulated when carbon is sequestered in the soil. This is assumed because there is lack of knowledge on this process. However, when such large amounts of carbon are sequestered in the soil, it could have an effect on the nitrogen accumulation in soil organic matter. Emission reduction measures and GHG abatement It is assumed that the emission reduction measures in the progressive scenario could be applied on the whole agricultural area from the beginning of the timeframe assessed, and that these measures will not affect yield levels negatively. The emission and sequestration factors of soil organic carbon are based on the default values of the IPCC, which provide only a very coarse estimation of the effect of measures such as reduced tillage and increased organic carbon input on the SOC content, and provide no quantitative information on the amount of organic inputs, effects on achieved yields, and the relation with nitrogen oxide emissions. The avoided emissions of substituting gasoline with bioethanol are derived from the well to wheel analysis of the JRC (JEC 2008). It is assumed that straw is not used for energy purposes and the assumed conversion efficiency of lignocellulosic bioethanol is rather conservative considering future learning and electricity production (Hamelinck and Faaij 2006).

6.4.3 Data requirements Spatiotemporal land use and GHG modelling requires numerous statistical and spatial data inputs. There is a lack of (spatial) data gathered and processed for Ukraine. There are many inconsistencies between several sources of statistical data: statistical data does not match spatial data, and spatial data have several quality issues related to resolution, classification and consistency. This is especially true for spatial data on pastures and abandoned agricultural land as most land use maps do not categorize these land use classes. In addition, data of several sources are combined, which are not necessarily 237

SHADES OF GREEN

consistent with each other. Examples are the data on agro-ecological suitability with the data on SOC, C:N ratio, and climate data. However, it is assumed that the currently available data has sufficient accuracy to distinguish patterns and hotspots of land use change and GHG emissions. However, if the presented approach is to be used for the exante assessment, monitoring and certification of biomass production, better data, and a finer modelling resolution is required.

6.5 Conclusions This study has analysed how the bioenergy potential and total greenhouse gas (GHG) balances in Ukraine may develop over time, taking into account development and emissions of total agricultural land use. The development in land requirements for food and feed production was analyzed spatially explicitly on an annual basis making use of the PCRaster Land Use Change (PLUC) model, adapted for characteristics of Ukraine. In the progressive scenario, 32.1 Mha of land could become available for energy crop production 2 by 2030. The spatiotemporal GHG module produced spatially explicit maps (1 km ) of the individual GHG emissions on an annual basis. The results show that the total cumulative GHG balance of -0.8 GT CO2-eq when wheat for ethanol is cultivated on the abandoned land, and of -3.8 GT CO2-eq when switchgrass is cultivated in the timeframe 2010-2030 (progressive scenario). If measures are taken to reduce agricultural CO2 and N2O emissions, the cumulative GHG emission avoidance by 2030 could even increase to 2.6 GT CO2-eq for wheat and 5.0 GT CO2-eq for switchgrass. When the available land is used for the re-growth of natural vegetation, a considerable amount of carbon is accumulated in the form of biomass and soil organic carbon. This could reach 4.3 GT CO2-eq in 2030. However, in the long run emission reductions decrease when forest becomes mature, whereas new bioenergy crops could continue to avoid GHG emissions in the future. For Ukraine, the GHG balance is dominated by the CO2-emisisons from carbon stock changes. This is due to the small ratio between the land in use for food and feed and the land that becomes available for bioenergy crop production. Therefore, a large area is converted to energy crops compared to the area of agricultural land that is intensified. In contrast, in countries that have already a more intensified agricultural production and a considerable livestock sector, the N2O emissions have a more significant contribution to the total GHG balance. The direct and indirect N2O emissions have been assessed in detail taking into account the spatial variation in crop requirements, nitrogen losses due to volatisation, leaching and run off, and by using land use, soil type, and nitrogen source specific N2O emission factors. The discriminating emission factors result in high spatial variation in N2O emissions. Incorporating volatisation, leaching and run offs, allows for evaluation of other 238

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

environmental impacts such as contamination of ground and surface water. However, as the carbon stock related emissions contribute primarily to the overall GHG emissions, more attention could be paid to refine modelling of the carbon stock changes. Both carbon and nitrogen related emissions could be modelled more accurately when the model is more process based instead of making use of default fractions and factors, and when better data is available. However, this requires more understanding of the underlying processes and the alterations of these processes in different conditions. Despite the uncertainties, this analysis shows that it is possible to link a spatiotemporal land use model with a dynamic GHG emission model and to assess spatial differentiations in GHG emission resulting from changes in land use and in land use management related to the implementation of bioenergy crop production. This is a large step forward compared to static and spatially aggregated GHG models and to models that have tried to calculate GHG footprints on aggregated spatial levels. The model allows for assessments of the best areas for intensification of agriculture and the best suitable areas for bioenergy crop production and it could contribute to assess under what conditions the GHG balance could be optimised in conjunction with safeguarding sufficient food and feed production.

239

SHADES OF GREEN

6.6 Appendix: input data GHG emissions and additional results 6.6.1 Input data Table 6.4: Input parameters for calculations of changes in carbon stock for the BAU en progressive scenario and the progressive scenario with mitigation measures. Parameter Maximum yield Weighted yield crops

Residue removal SOC Land use factor

BAU

PROG

Mitigation

ton dm/ha/yr

2.29

2.33

6.04

6.04

ton dm/ha/yr

1.94

1.94

4.57

4.57

Wheat (energy)

ton dm/ha/yr

9.1

9.1

9.1

9.1

Switchgrass (energy)

ton dm/ha/yr

21.4

21.4

21.4

21.4

Forest re-growth

ton dm/ha/yr

7.1

7.1

7.1

7.1

Cereals

fraction

0.5

0.5

0.5

0.25

Other crops

fraction

0.0

0.0

0.0

0.0

Cold temperate moist

cropland

0.72

0.72

0.69

0.69

pasture

1.00

1.00

1.00

1.00

wheat

0.69

0.69

0.69

0.69

switchgrass

1.00

1.00

1.00

1.00

Cold temperate moist

Cold - Warm temperate dry

SOC input levels

Current

Yield pasture

Cold - Warm temperate dry

SOC Management factor

Unit

Cold temperate moist

Cold - warm temperate dry

cropland

0.83

0.83

0.80

0.80

pasture

1.00

1.00

1.00

1.00

wheat

0.80

0.80

0.80

0.80

switchgrass

1.00

1.00

1.00

1.00

cropland

1.08

1.08

1.00

1.08

pasture

0.95

0.95

1.14

1.14

wheat

1.00

1.00

1.00

1.08

switchgrass

1.14

1.14

1.14

1.14

cropland

1.02

1.02

1.00

1.02

pasture

0.95

0.95

1.14

1.14

wheat

1.00

1.00

1.00

1.02

switchgrass

1.14

1.14

1.14

1.14

cropland

0.92

0.92

1.00

1.44

pasture

1.00

1.00

1.00

1.11

wheat

1.00

1.00

1.00

1.44

switchgrass

1.00

1.00

1.00

1.11

cropland

0.95

0.95

1.00

1.37

pasture

1.00

1.00

1.00

1.11

wheat

1.00

1.00

1.00

1.37

switchgrass

1.00

1.00

1.00

1.11

240

6.

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine

Table 6.5: Emission factors of direct and indirect N2O emissions differentiated for sediment type, land use and nitrogen source. Clay

Peat

Source of Nitrogen

Grassland

Sand Arable

Grassland Arable

Grassland Arable

Nitrate based fertilizer

0.010

0.005

0.015

0.008

0.020

0.010

Ammonium based fertilizer

0.005

0.004

0.008

0.006

0.010

0.008

Fertilizer BAU and PROG a

0.008

0.005

0.011

0.007

0.015

0.009

Fertilizer in mitigation option

0.005

0.004

0.008

0.006

0.010

0.008

Pig slurry surface applied

0.005

0.008

0.008

0.011

0.010

0.015

Solid pig manure

0.002

0.003

0.003

0.004

0.003

0.005

Solid cattle manure

0.002

0.003

0.003

0.004

0.003

0.005

Poultry manure

0.002

0.003

0.003

0.004

0.003

0.005

Grazing

0.020

Other manure Total manure

0.002 b

0.002

0.030 0.003

0.003

0.003

0.003

0.040 0.004

0.003

0.004

0.004

0.005 0.005

Crop residues cereals

0.002

0.003

0.004

Crop residues vegetables

0.020

0.030

0.040

Crop residues other crops

0.010

0.015

0.020

Total residues

c

Atmospheric deposition

0.003 0.004

0.005

0.003

0.006

0.005

0.007 0.008

0.006

Mineralization

0.004

0.003

0.006

0.005

0.026

0.026

N from volatisation d

0.010

0.010

0.010

0.010

0.010

0.010

N from leaching and run off d

0.008

0.008

0.008

0.008

0.008

0.008

a

It is assumed that ammonium based and nitrate based fertilizes are used 50-50% in the BAU and progressive scenario. it is assumed that in scenario with mitigation measures, only ammonium based fertilizers are used. b Based on the animal manure composition c Based on the crop composition c Emission factors of indirect N2O emissions are based on the IPCC (2006)

241

SHADES OF GREEN

GJ/odt

GJ/ odt

kg CO2/odt

odt/ha

GJ/ha

Kg CO2/ha

38

50

0.65

17

11.05

553

9.1

101

5028

Max abatement

MJfuel/ MJbiomass

Max ethanol yield

gCO2 /MJfuel

Max yield

Energy content raw biomass c

gCO2 /MJfuel

Abatement per odt

Conversion factors a

Ethanol from biomass

Abatement (ex N2O) b

Wheat

Abatement a

Abatement

Table 6.6: Abatement figures for the replacement of gasoline by ethanol from switchgrass and wheat based on the figures provided by JRC.

Switch 71 78 0.43 18.4 7.91 617 21.4 169 13207 grass a Figures derived from JRC (JEC 2008).Assuming ‘wheat ethanol production with natural gas as a process fuel in a conventional boiler’ and ethanol production from farmed wood which is also applicable for perennial grasses. b Based on de Wit et al. (2011c). c Based on Bhoemel et al. (2008). d The maximum yield is derived from the Refuel study (de Wit and Faaij 2010; Fischer et al. 2010a; Fischer et al. 2010b).

6.6.2 Additional results In Figure 6.11, snapshots of annual nitrous oxide emissions (left), CO2 abatements (middle), and CO2 emissions (right) are shown for all agricultural land use and cultivation of switchgrass for ethanol on abandoned agricultural land for the years 2010, 2020 and 2 2030 for the progressive scenario. These are annual emissions in tons of N2O, CO2 per km . The emissions are therefore directly related to the change in land use and management of the specific year. Nitrous oxide emissions increase, especially in the areas where (low productive) agriculture land is abandoned and becomes in use for switchgrass cultivation. Because of the high yields, high nitrogen inputs are required, which results in higher emissions. At the same time, at the same locations large abatement potentials are achieved. The abatement potentials are directly related to the potential yield of switchgrass. CO2 emissions from cropland remaining cropland increase in the progressive scenario due to shift from reduced to full tillage. The SOC in pastures is expected to increase due to better management. High GHG emission mitigation potentials are achieved when abandoned cropland is used for switchgrass cultivation.

242

Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances in Ukraine 6.

Figure 6.11: Snapshots of annual N2O emissions (left), CO2 abatement (middle), and CO2 emissions (right) for all agricultural land use and cultivation of switchgrass for ethanol on abandoned agricultural land for the years 2010, 2020 and 2030 for the progressive scenario.

243

7.

Summary and conclusions

7

Summary and Conclusions

245

SHADES OF GREEN

7.1 Research context All societies require energy services to meet basic human needs and to facilitate productive processes (IPCC 2011). Currently, 85% of the total primary energy supply originates from fossil fuels (IEA 2010a; IPCC 2011). However, the fossil fuel dominated energy system cannot be sustained in the long run because of the finity of fossil resources, the unequal distribution of resources, and the major contribution of the use of fossil fuels to anthropogenic GHG emissions. An increased deployment of renewable energy resources to substitute fossil fuels will be required to make the energy system more sustainable (IPCC 2007a). Modern biomass applications contributed worldwide 11 EJ in 2008, and are expected to play an important role in future energy supply (IEA and OECD 2011; IPCC 2011). For example, based on scenario analyses, the Intergovernmental Panel on Climate Change (IPCC) estimates that a biomass deployment level of 120-190 EJ in 2050 is required in order to meet the GHG mitigation goals related to a stabilization of atmospheric CO2-eq concentration at a level of less than 440 ppm by 2100 (IPCC 2011). However, the increase in production and use of bioenergy is not only driven by its GHG mitigation potential (if produced sustainably), but also by the relatively easy implementation of bioenergy in existing energy infrastructure, the versatility of biomass as a resource, the diversification of energy supply and subsequent increase in energy security, the potential contribution to rural development and the potential restoration of degraded lands. However, an increased implementation of dedicated bioenergy crop production could have significant adverse socio-economic and environmental impacts such as deforestation, loss of carbon sinks, biodiversity and other ecosystem functions and services, displacement of people, and an increased competition for land, water and other production factors, which in turn can lead to higher food prices (IPCC 2011). Many of these impacts are related to land use change (LUC) (Wicke et al. 2012). In order to achieve a high deployment level of bioenergy, competition between food, feed and fuels - and therefore also indirect land use change (iLUC) - need to be avoided. This is possible by balancing the increased production of biomass for energy with improvements in agricultural management (Dornburg et al. 2010; Wicke et al. 2012). Furthermore, the key environmental concerns should be addressed by selecting appropriate bioenergy systems and applying adequate land use planning (Dornburg et al. 2010). The impacts and performance of biomass production and use are region- and site-specific (IPCC 2011) and impacts occur at local, regional as well as global level (van Dam et al. 2010b). Implementation of effective sustainability frameworks, for example, by developing

246

7.

Summary and conclusions

certification schemes, could mitigate negative environmental impacts and allows simultaneously for contributing to multiple objectives of sustainable development. At several levels, initiatives for sustainability criteria, codes of conduct and protocols have been and are currently developed to deal with these sustainability issues. At present, a key bottleneck for both market players and government is how such criteria can be met in practice and how impacts can be quantified in a verifiable and reliable manner. Strong improvement in spatially explicit potential and impact analyses are required to allow for effective certification, sound planning of sustainable investments in the future and good governance of land use and the agricultural sector, all in direct relation to increasing biomass production and subsequent use for energy and materials.

7.2 Aim and research questions This thesis aimed to examine how potentials, costs, and environmental impacts of bioenergy production can be assessed, taking into account the avoidance of ilUC and the spatiotemporal variability of the biophysical and socio-economic context. To this end, the following research questions were addressed: I. How can potential land availability for energy crops be assessed spatially and temporal explicitly, given that iLUC should be avoided and therefore taking into account the development in other land use functions? II. How can the economic viability (the location specific competition with other agricultural land uses, the cost of biomass feedstock production and the logistics of the supply chain) and the environmental impacts of bioenergy production (impacts on GHG emissions, soil, water and biodiversity) be assessed spatially and temporally explicitly. III. What are the potentials, economic performances and environmental impacts of bioenergy production in different settings? IV. What reliability can be obtained using the data available and the methods developed in this study? The research questions are covered in Chapters 2 through 6. In Chapter 2 and 3, the economic viability and the potential environmental impacts of regional bioenergy production chains was assessed spatially explicitly, taking into account spatial variation in agro-ecological suitability, current land use and other biophysical factors. The GIS-based methodologies developed in these chapters allow for spatial explicit but static assessments of costs and environmental impacts. The north of the Netherlands was selected as a case study area because of the high competition for land and related intensive land use, and because of the proper availability of detailed spatial data. In Chapter 4, a new land use change model (PLUC) was developed to assess the land 247

SHADES OF GREEN

availability for bioenergy crops, taking into account country specific land use change drivers. The developed model enables spatial and temporal explicit assessment of land availability for bioenergy crops and facilitates spatiotemporal uncertainty assessment based on stochastic input parameters. In Chapter 5, the developments in costs of biomass feedstock and the logistics of bioenergy supply chains were analysed, given the developments in potential land availability and technological learning (improvements in yield and conversion technologies). The coupling of the spatiotemporal land use model, the spatially explicit costs of biomass production and logistics, and the temporal cost developments, allowed for spatiotemporal assessment of bioenergy supply costs. In Chapter 6, the developed land use change model was adapted for Ukraine and extended with a GHG emission module, in order to analyse the potential developments in the land use change related GHG emissions, including the implementation of bioenergy crops and intensification of the agricultural sector. The link between the dynamic GHG emission calculation module and the land use change model allows for spatiotemporal GHG impact assessment of total land use change. Mozambique and Ukraine were selected as case study areas because of their high bioenergy production potential related to the low population density and the favourable climate for biomass production, and because of the diversity of the environmental and socio-economic conditions of the countries. In chapters 2, 3, 5, and 6, the performance of typical crops for production of first and second generation biofuels is analysed. In (almost) all case studies, these energy crops are assumed to be used for ethanol production in order to allow for comparison of the potentials, costs, and environmental impacts for typical first and second generation biofuel options. The complexity and level of integration of methodologies developed in this thesis increase from chapter to chapter and evolves from spatially explicit and static modelling (in Chapter 2 economic performance and in Chapter 3 environmental impacts) to spatiotemporal and dynamic modelling (in Chapter 4 land use change and in Chapter 5 cost supply development). In Chapter 6, the spatiotemporal land use modelling is integrated with dynamic GHG emission modelling, which allows for spatiotemporal integrated impact assessment. In all chapters, the limitations of the methods and the data used in the assessments are discussed. Table 7.1 presents an overview of the chapters and the research questions addressed.

248

7.

Summary and conclusions

Table 7.1: Overview of the settings of the thesis chapters and the research questions addressed in them. Chapter 2 3 4 5 6

Potential, spatial distribution and economic performance of regional biomass chains Spatial variation of environmental impacts of regional biomass chains Spatiotemporal land use modelling to assess land availability for energy crops Spatiotemporal cost-supply curves for bioenergy production Integrated spatiotemporal analysis of agricultural land use, bioenergy production potentials and related GHG balances

Research questions I II III

IV

































7.3 Summary of the results Chapter 2 addresses research question II, III and IV by analysing the spatial variation in economic viability of ethanol production from Miscanthus and sugar beet in the north of the Netherlands. The competitiveness of bioenergy crops was assessed by comparing the Net Present Value (NPV) of perennial crops; current rotations; and rotation schemes, which include additional years of sugar beet; and by comparing the production cost of bioethanol with average petrol prices. The current land use and soil suitability for present and bioenergy crops were mapped using a Geographical Information System (GIS,) and the spatial distribution of economic profitability was used to indicate where land use change is most likely to occur. The productions costs of bioethanol comprise costs associated with cultivation, harvest, transport and conversion to ethanol. The NPVs and cost of feedstock production were calculated for seven soil suitability classes. The results show a high spatial variation of both the cost of biomass feedstock production and the profitability of biomass production compared to conventional agricultural crops. At current market prices, bioenergy crops are not competitive with conventional cropping systems on soils classified as ‘suitable’. On less suitable soils, the return on intensively managed crops is low and perennial crops achieve better NPVs than common rotations. The results showed that minimum feedstock production costs are 5.4 €/GJ for Miscanthus and 9.7 €/GJ for sugar beet depending on soil suitability. Ethanol from Miscanthus (24 €/GJ) is a better option than ethanol from sugar beet (27 €/GJ) in terms of costs, but bioethanol production from domestically cultivated crops is not competitive with petrol (12.3 €/GJ) production under current circumstances. However, when oil prices increase, or when technological learning and biorefining results in more cost effective supply chains, the competitiveness could improve. This chapter provides a generic methodology to identify promising locations for bioenergy crop production from an economic point of view, taking

249

SHADES OF GREEN

into account the spatial variation in current land use and biophysical factors such as soil properties and water availability. Chapter 3 addresses research questions II, III and IV. In this chapter, the spatial variation of potential environmental impacts of bioenergy crop production is quantitatively assessed. The cultivation of sugar beet and Miscanthus for bioethanol production in the North of the Netherlands is used as a case study. The environmental impacts included in this study are: greenhouse gas (GHG) emissions (during lifecycle and related to direct land use change), soil quality, water quantity and quality, and biodiversity. For each impact, suitable indicators and methods were selected based on extensive literature review, an adapted where needed in order to use them spatially explicitly. The spatial variation in environmental impacts related to the spatial heterogeneity of the physical context is assessed using a Geographical Information System (GIS). The case study shows that there are large spatial variations in environmental impacts of the introduction of bioenergy crops. In general, sugar beet causes relatively many negative environmental impacts especially in pasture areas. In these areas, the LUC related GHG emissions are up to 148 kg -1 -1 CO2-eq GJethanol and the risk on soil erosion increases to 9 ton ha . Also, there is a high risk of biodiversity loss in these areas. Positive impacts in these areas are a decrease of 75 -1 mg l in the NO3 concentration of groundwater and a decrease in the seasonal water deficits with 100 mm. When arable land is converted to Miscanthus, GHG emissions can -1 be reduced by -159 kg CO2-eq GJethanol , the risk on soil erosion could be reduced with 4 -1 -1 ton ha , the NO3 concentration is reduced with 53 mg l , it has a positive effect to biodiversity, but on the other hand, seasonal water depletion can increases with 150 mm. For both crops, the western wet pasture areas appear to be the area with most negative impacts. The spatially combined results of the environmental impacts illustrate that there are several tradeoffs between environmental impacts: there are no areas were no negative environmental impacts occur. The assessment demonstrates a framework to identify areas with potential negative environmental impacts of bioenergy crop production and areas where bioenergy crop production have limited or positive environmental impacts. Chapter 4 addresses research question I, III and IV by developing a model to assess future developments in land availability for bioenergy crops in a spatially explicit way, while taking into account the prevention of competition with other land use functions, and therefore the avoidance of iLUC. The country specific land use change drivers for the development in other land use functions, such as land for food, livestock and material production and the uncertainties in these drivers are identified and included in the model. This spatiotemporal PCRaster based land use change (PLUC) model is demonstrated with a case study on the developments in the land availability for bioenergy crops in 250

7.

Summary and conclusions

Mozambique in the timeframe 2005-2030. The developments in the main drivers for agricultural land use (i.e. demand for food, animal products and materials) were assessed, based on the projected developments in population, diet, GDP and self-sufficiency ratio. Two scenarios were developed: a business-as-usual (BAU) scenario and a progressive scenario. Land allocation was based on land use class-specific sets of suitability factors. 2 The land use change dynamics were mapped on a 1 km grid level for each individual year up to 2030. In the BAU scenario, 7.7 Mha and in the progressive scenario 16.4 Mha could become available for bioenergy crop production in 2030. Based on the Monte Carlo analysis, a 95% confidence interval of the amount of land available and the spatially explicit probability of available land was found. The bottom-up approach, the number of dynamic land uses, the diverse portfolio of land use change drivers and suitability factors, and the possibility to model uncertainty allows for significantly improved integral modelling of land availability for bioenergy potentials. In addition, it provides opportunities to explore the preconditions for high bioenergy deployment levels while avoiding iLUC. Chapter 5 addresses research questions II , III and IV by assessing how bioenergy cost and supply potential develop over time in a spatially explicit way. The assessment of the cost and supply potential is based on the developments in land availability, the suitability of the land that is and can become available, the disaggregated cost break down of energy crop production, the transportation distance of feedstock to conversion plant, the cost of conversion, the transportation distance from plant to harbour and the cost of international shipping. The supply chains of eucalyptus (torrefied) pellets and sugarcane ethanol in Mozambique are used as a case study. The developments in land availability for energy crops in Mozambique are based on the findings of chapter 4. The results of chapter 5 show a large spatial variation in supply chain costs, which is the result of spatial variation in feedstock production costs, and primary and secondary transportation costs. Most promising areas for eucalyptus and sugarcane production are scattered in the central south, the central, and the north eastern part of Mozambique where agro-ecological conditions are relatively favourable, where sufficient feedstock can be produced to meet the input requirements of the conversion plant, and where infrastructure is available. In the progressive scenario, the total calculated potential for eucalyptus pellet production amounts 3200 PJ in 2030, of which 2500 PJ could be exported to Europe below a market price level of 8 €/GJ; for sugarcane ethanol the total potential amounts 850 PJ of which 500 PJ could be exported below a price level of 30 €/GJ. The location of production is the key factor for cost effective production. This is especially true for countries with a high heterogeneity in agro-ecological suitability and with a low availability of infrastructure. When road and railroad infrastructure would be improved, cost of logistics could be reduced and more areas could become economically viable for bioenergy production. This 251

SHADES OF GREEN

study demonstrates an approach which enables the assessment of the development of bioenergy potentials and costs over time in a spatially explicit way. As environmental and socio-economic impacts of bioenergy supply chains are highly related to the biophysical and socio-economic context of the location of production, a spatially explicit assessment of bioenergy production potential is a suited approach for the design, optimisation and analysis of the impacts and sustainability of bioenergy chains. Chapter 6 addresses research question I, II, III and IV by analysing spatially explicitly how bioenergy potential and greenhouse gas (GHG) emission reduction in Ukraine can develop in the timeframe 2010-2030, while taking into account development and emissions of other agricultural land use functions. The development in land requirements for food and feed production is analysed spatially on an annual basis making use of the PCRaster Land Use Change (PLUC) model. The model was tailored for Ukraine, by adapting the dynamic land use classes and suitability factors and their characteristics for the Ukrainian situation. Two scenarios for the period 2010-2030 have been assessed: a Business As Usual scenario (BAU), in which current trends in productivity are continued; and a progressive scenario, which projects a convergence of yield levels for Ukraine with the level of West European countries. In the progressive scenario, 32.1 Mha land could become available for energy crop production by 2030. The abandoned areas are mainly situated in the north, east and southern parts of Ukraine. The projected land use developments serve as input for the spatiotemporal GHG balance, which includes CO2, N2O and CH4 emissions related to changes in management and land use, as well as the abatement of GHG emissions by replacing fossil fuels by assumed bioethanol production from wheat and switchgrass. The 2 spatiotemporal GHG module produces spatially explicit maps (1 km resolution) of the individual GHG emissions on an annual basis. High carbon sequestration levels are achieved when switchgrass is cultivated or when natural vegetation regenerates on abandoned agricultural lands. Total N2O emissions increase when the abandoned agricultural land is cultivated with switchgrass or wheat. Especially in areas where dedicated cropland is expanding at the expense of mosaic cropland-pasture areas. The results show that a total cumulative GHG balance of -0.8 GT CO2-eq for wheat and -3.8 GT CO2-eq for switchgrass could be achieved in 2030 in the progressive scenario. The negative balance of wheat is also caused by the re-growth of natural vegetation of abandoned agricultural areas that are excluded for bioenergy crop production. When additional GHG mitigation measures are taken (such as reduced tillage, increased organic carbon input and use of improved fertilisers), in order to reduce agricultural CO2 and N2O emissions, the cumulative GHG balance could even increase to -2.6 GT CO2-eq for wheat and -5.0 GT CO2eq for switchgrass by 2030. When the available land is used for the re-growth of natural vegetation, a considerable amount of carbon will be accumulated in the form of biomass and soil organic carbon. This could reach -4.4 GT CO2-eq in 2030. However, this carbon 252

7.

Summary and conclusions

sequestration can only be obtained once in contrast to the abatement of bioenergy crops. Considering the progressive scenario and a timeframe up to 2100, the total cumulative GHG mitigation is estimated at ± -5.5 GT CO2-eq for re-growth of natural vegetation, -6 GT CO2-eq for wheat and -15 GT CO2-eq for switchgrass. The spatiotemporal GHG module in conjunction with the PLUC model allows for spatially explicit and dynamic modelling of GHG emissions resulting from changes in land use and management related to the implementation of bioenergy crop production. Based on the methodology development and findings of chapter 2 – 6, answers to the main research questions and recommendations for policy and further research are given.

7.4 Main findings and conclusions I.

How can the development in potential land availability for energy crops be assessed spatially and temporal explicitly, taking into account the development in other land use functions and given that iLUC should be avoided?

In this study a new LUC model was developed (PLUC) to assess the development in land 2 availability for bioenergy crops on a detailed spatial level (1 km ), taking into account the dynamics and uncertainties of key drivers of LUC. The main drivers for the development in demand for agricultural products in a region are development in population, GDP, food intake and material use per capita and SSR. The amount of land required to meet the total demand, depends on the efficiency of the agricultural sector and the agro-ecological suitability of the area of production for specific land uses (e.g. cropland, pasture, forest). Since it is uncertain how LUC drivers evolve, a scenario approach can be used to explore potential long-term developments in LUC driving forces. The allocation of land to dynamic land uses classes is based on the suitability of the location for a specific land use class. This is defined by a combination of spatially explicit suitability factors related to the agro-ecological suitability, the accessibility, conversion elasticity and neighbourhood. For each suitability factor, the direction, type and extent of correlation need to be determined. The suitability factors, their characteristics and their relative importance are region and land use class specific. Areas that are not suitable (e.g. steep slopes) or not allowed to be converted to agricultural land (e.g. conservation areas), need to be excluded. Starting point of land use allocation is the map of current land use calibrated with the statistics on demand for food feed and material, the agricultural productivity, and the distribution of the land use classes over the agro-ecological suitability. Land is allocated to a land use class in time 253

SHADES OF GREEN

steps of one year. The total allocation is completed for one time step when all the land use classes have been allocated and the production of these land use classes meets the total demand for that particular year. The modelling comprises a feedback loop: the land use map resulting from the allocation in time step t serves as input for the allocation in time step t+1. The output of this type of dynamic land use change modelling are maps of land 2 use on a detailed spatial level (1km ) and the land availability for bioenergy crops for each year within the modelling period. The major advantage of this model framework is its ability to deal with stochastic input data. This enables spatio-temporal Monte Carlo (MC) runs that evaluate uncertainty propagation. PLUC can stochastically model time series (e.g. crop demand and agricultural productivity), spatial input parameters (e.g. population density and productivity), and characteristics of suitability factors (e.g. the maximum distance of effect in the distance to road). The stochastic modelling enables sensitivity analyses of the results for uncertainties in key parameters. This results in maps depicting the probability of the availability of land 2 at a specific location (grid cell, 1 km ) in a specific time step (year). The land use model developed in this study is an advanced tool to assess future land use dynamics and land availability for bioenergy crops. Applying a scenario approach on the key drivers of land use change and using a food first paradigm, allows for an evaluation of the biomass potentials that can be achieved without competition with food and feed, and the required conditions to realize these potentials. The bottom-up approach, the number of dynamic land uses, the diverse portfolio of LUC drivers and suitability factors, and the possibility to model uncertainty is a step forward in modelling the land availability for energy crops. As biomass yields, production costs, logistics, and environmental impacts are strongly related to location specific biophysical conditions (e.g. agro-ecological suitability, availability of infrastructure, soil properties, climate conditions etc); spatially explicit assessment of land availability for bioenergy crops is an important precondition to design bioenergy supply chains and logistics and assess bioenergy production potential and environmental and socio-economic impacts. The model has now been tailored to, and demonstrated for Mozambique and Ukraine. Still, it is a flexible model which can be used for other countries or regions when input data, allocation rules and characteristics of suitability factors are adapted. II.

How can the economic viability in terms of production costs and competitive advantage and the environmental impacts e.g. impacts on GHG emissions, soil, water and biodiversity of bioenergy production be assessed spatially and temporal explicitly?

254

7.

Summary and conclusions

In an early stage of this study, methods were environmental performance in a spatially explicit variability in biophysical context. In a later stage, a for the assessment of economic viability and the account both spatial and temporal variation.

developed to assess economic and way, taking into account the spatial spatiotemporal model was developed GHG emissions in order to take into

The economic viability of bioenergy production depends on the competitive advantage op bioenergy crop production compared to other land uses and the competitiveness of bioenergy production compared to the reference energy system. Three key cost factors determine the cost of bioenergy production: cost of biomass feedstock production, the costs of supply chain logistics and the cost and efficiency of the conversion technology. Feedstock production costs can be assessed calculating the net present value (NPV) of all costs items (land, labour, machinery, inputs) and the biomass yield during the lifetime of the biomass plantation. The spatial variation in yield and related costs and the competitiveness with reference land use can be calculated by combining the land use map and the crop specific agro-ecological suitability maps. The spatial variation in cost of biomass logistics can be analysed making use of data on the scale of the conversion plant and the spatial information on land availability, yield levels, and the availability and the quality of infrastructure. The conversion costs comprise investment costs, operation and maintenance (O&M) costs, and energy input costs. The costs of biomass production and conversion change over time due to technological learning. Integrating the projections on technological learning in the cost calculations, interlinked with the spatiotemporal land use modelling, allows for the spatial and temporal explicit evaluation of the economic performance of bioenergy supply chains. The selection of environmental impacts of bioenergy production analysed in this thesis was based on the areas of concern identified by several national and international initiatives regarding sustainability criteria of bioenergy production (EC 2009a; NEN 2009; RSB 2010). The environmental impacts included are the GHG emissions (during lifecycle and related to land use change) and the impacts on soil, water, and biodiversity. The environmental impacts were quantitatively and spatially explicitly assessed by developing /adapting existing methodologies for spatial detailed analysis. The impacts on water were assessed making use of a simple water balance and taking into account detailed spatial data on land use change, effective precipitation and potential evapotranspiration. The impacts on soil were assessed making use of the Wind erosion Equation (WEQ) and accounting for the spatial variation in land use change, vegetation characteristics, soil characteristics, and wind, precipitation and temperature. The impacts on biodiversity were explored using the Mean Species Abundance (MSA) and the High Nature Value

255

SHADES OF GREEN

(HNV) indicators changes in land use and management and the distribution of natural areas and endangered species. In an early stage, the Miterra model (Lesschen 2008; Velthof et al. 2009) was used to assess the GHG emissions and the impact on water quality related to land use change. This deterministic, static and vector based model simulates the nitrogen and phosphorous balance, emissions of NH3, N2O, NOx, and CH4, leaching of N, NO3 concentration in groundwater, and changes in soil and biomass carbon stocks related to changes in land use and management. In a later stage, a new model was developed in order to assess GHG 2 emissions in dynamically on a spatial detailed level (1 km ). The calculation of the nitrogen pathways was primarily based on the methods developed by Lesschen (2008) and Velthof et al. (2009) but adapted for raster based and dynamic calculations. In addition, instead of the default N2O emission factor proposed by the IPCC, nitrogen source and soil type and land use specific N2O emission factors developed by Lesschen et al. (2011) were implemented in the model and the leaching and runoff factors were adapted. The calculations of carbon emissions are based on the method proposed by the IPCC (2006) and are modelled dynamically. By integrating this GHG emission module to the spatiotemporal land use model, the GHG emissions can be calculated spatial and temporal explicitly. This new model is able to account for spatial variation in land use, yield level, soil characteristics, climate, slope; and temporal variation in land use, management, and yield levels. In addition, it allows for assessments of the best areas for intensification of agriculture and the most suitable areas for bioenergy crop production. Moreover, it could contribute to assess under what conditions the GHG balance could be optimised in conjunction with safeguarding sufficient food and feed production. The methods and models developed provide an approach to identify ex ante the areas where implementation of bioenergy production is or could become economically viable and the areas with little negative or even positive environmental impacts. Comparing the spatial distribution of the several environmental impacts and the economic viability allows for the evaluation of trade-offs between environmental impacts and between environmental and economic performance. The integration of these models and methods enables the identification of go and no-go areas for bioenergy production from an economic and environmental point of view and when competition for land is to be avoided. III.

What are the potentials, economic performance and environmental impacts of bioenergy production in different settings?

When ilUC needs to be avoided, the amount of bioenergy that can be produced depends on the developments in demand of other agricultural products, the rate of intensification 256

7.

Summary and conclusions

of the agricultural sector, and the suitability of the land that becomes available for energy crop production. In the Netherlands, population is increasing at a low pace and dietary intake is assumed to stabilise at current levels. If it is assumed that the self sufficiency ratio remains stable, the total demand for food and feed products is expected to increase slightly over time. The Netherlands has one of Europe’s most efficient and technologically advanced agricultural sectors (de Wit et al. 2011a). Therefore, the yield gaps and related opportunities for efficiency improvements are relatively limited (especially compared to Ukraine and Mozambique). This results in relatively low land availability for bioenergy crops ranging of 41 to 51 kha in the 3 northern provinces in 2030 (de Wit and Faaij 2010; Fischer et al. 2010a; Fischer et al. 2010b). Considering a regional average yield level of Miscanthus and sugar beet, 6 to 9 PJ ethanol could be produced in Groningen, Friesland and Drenthe by 2030. As it was not assessed where land will become available, no precise estimation can be made on the bioenergy crop production potential. The cost of ethanol production, based on the least feedstock production costs, are 24 €/GJ for Miscanthus and 27 €/GJ for sugar beet ethanol (considering the performance of conversion technologies available on the short term). In Mozambique, the demand for agricultural products will increase due to a strongly increasing population and improvements in dietary intake. On the other hand, as the current agricultural productivity is very low, it offers a large potential for improvement. When a Business as usual (BAU) scenario is assumed, in which current trends in agricultural productivity are continued, 7.7 Mha could become available; In a progressive scenario, in which a agricultural productivity is increased considerably, 16.4 Mha could become available for bioenergy crop production in 2030. When the available land is used for eucalyptus 1340 PJ of torrefied pellets could be produced in the BAU scenario in 2030 and in 3200 PJ in the progressive scenario. When the available land is used to cultivate sugar cane for ethanol production, 350 PJethanol in the BAU and 850 PJethanol in the progressive scenario could be produced in 2030. The least cost of torrefied pellets are 5 €/GJ and for ethanol from sugar cane the least costs are 14 €/GJ. The least cost areas, are those areas that are high productive, where sufficient feedstock can be produced to meet the input requirements of the conversion plant, and that are in the vicinity of infrastructure and harbours. When road and railroad infrastructure are improved, cost of logistics can be reduced and more areas can become economically viable for bioenergy production. In Ukraine, the demand for agricultural products is expected to increase slightly. The population size decreases and the dietary intake is stabilizing, but Ukraine is expected to 257

SHADES OF GREEN

increase their export levels. Despite the favourable agro-ecological conditions, the productivity of the agricultural sector is low, which offers large potentials for improvement by exploiting the yield gaps. In the BAU scenario, in which little to no improvements in agricultural are made, limited land (0.03 Mha) is available for bioenergy production. However, in the progressive scenario 32.1 Mha could become available for energy crop production by 2030. When this land is used for ethanol production from switchgrass, 3 PJethanol in the BAU and 2230 PJethanol in the progressive scenario can be produced in 2030. When wheat is cultivated on the abandoned land, 1 to 2370 PJethanol can be produced in 2030. Based on the least feedstock cost in Ukraine provided by de Wit and Faaij (2010), the ethanol production costs are calculated and range between 9 to 11 €/GJ. Because Ukraine consist predominantly of agricultural land (79 %), and the majority of the remainder land is either forest (15%) or static land uses (5%, e.g. conservation areas, build environment), there is almost no land available for energy crop production unless the agricultural land required for food and livestock decreases by means of more efficient production. For Mozambique, there is currently relatively large land availability (9 Mha). Besides the current agricultural land use (20%), forest (60%) and static land uses (9%), approximately 10% of the land could be used for bioenergy crops. However, the land availability for bioenergy crops and forest areas decrease when land required for food and feed expands due to an increasing demand and when limited up to no improvements in the agricultural efficiency are achieved. Due to the major differences in biophysical characteristics, socio-economic conditions, historical developments, stage of development and policy specific context, the land use change dynamics in Mozambique and Ukraine are determined by different drivers, or some drivers play a more or less important role. Because of the large dependence on subsistence farming, the lack of infrastructure and the dependence on local markets, agricultural land in Mozambique is expected to concentrate in the more densely populated areas, close to major cities and close to the road network. In Ukraine, less people depend directly on agriculture, a relatively dense road network is available, and a gradual shift towards larger, more commercial farms is ongoing. Here, the agro-ecological suitability is the key factor for the location of agricultural land and a smaller correlation is observed with population density or accessibility. Because of these major differences, the parameters in the PLUC model (drivers, allocation rules, suitability factors and their characteristics and relative importance), need to be adapted for every setting. The assessment of the development in cost of the supply chain shows that costs can be significantly reduced by means of yield improvements, improved logistics, and improved pre treatment and conversion technologies. 258

7.

Summary and conclusions

The integrated assessment of environmental impacts in the north of the Netherlands shows a high spatial variation in impacts. There are no areas where only positive effects occur when bioenergy crops are introduced and there are tradeoffs between the impacts. The case study shows that there are large spatial variations in environmental impacts of the introduction of bioenergy crops. In general, sugar beet causes relatively many negative environmental impacts especially in pasture areas. In these areas, the LUC related GHG -1 emissions can be up to 148 kg CO2-eq GJethanol and the risk on soil erosion can increases to -1 9 ton ha . Also, there is a high risk on biodiversity loss in these areas. Positive impacts in -1 these areas are a potential decrease of 75 mg l in the NO3 concentration of groundwater and a possible decrease in the seasonal water deficits of 100 mm. When arable land is -1 converted to Miscanthus GHG emissions can be reduced by -159 kg CO2-eq GJethanol , the -1 risk on soil erosion can be reduced with 4 ton ha , the NO3 concentration is potentially -1 reduced with 53 mg l , it has a positive effect to biodiversity, but seasonal water depletion can increases with 150 mm. The spatial patterns of impacts are mostly related to the spatial pattern of current land use; i.e. the magnitude of the environmental impact is largely influenced by the type of land use that is converted to energy crops. The integrated analysis of GHG balance of bioenergy crop production and the intensification of the agricultural sector shows that high GHG emission reduction potentials can be achieved when the agricultural land use is intensified and bioenergy crops are implemented. Large emission reductions are only achieved when the entire agricultural sector is managed more sustainably. Global comparison of a number of quantitative results In Table 7.2 an overview of the potentials, cost and environmental impacts of bioenergy production is provided for the different geographical settings. It can be concluded that in the progressive scenarios considerable amounts of land may become available for energy crop production in Mozambique and in Ukraine, without conflicting with other land uses. And that this could potentially result in high GHG emission savings. However, the low and decreasing land potential land availability in the BAU scenario indicates a higher competition for land in the future. Therefore, it must be stressed that a large-scale sustainable bioenergy sector can only be established if it is developed simultaneously with a more productive and sustainable agricultural sector. This implies a discontinuation of current trends. For Mozambique this means a shift away from subsistence towards commercial farming, and from pastoral towards mixed livestock systems. This requires changes in agricultural management (especially deployment of fertilizer and improved seeds), development of regional or (inter-) national markets, improved logistics, training and better overall capacities and governance of the 259

SHADES OF GREEN

agricultural sector. For Ukraine, this would imply land reforming that facilitate commercial farming, alignment with international standards on agricultural products, an economic system that allows for off- farm income and access to capital and agricultural inputs. Table 7.2: Overview of the potentials, cost and environmental impacts of bioenergy production is provided for the different geographical settings.

Land availability

a

Bioethanol production potential 1st gen ethanol b Bioenergy production potential from woody/herbaceous crop c st

Ethanol costs 1 gen ethanol

d

Bioenergy costs 2nd gen ethanol / torrefied pellets d

Unit

The Netherlands

Mozambique

Ukraine

Mha

0.04 - 0.05

7.7 -16.4

0.03 - 32.1

PJ

6-7

350 - 850

1 - 2370

PJ

7-9

1340 - 3200

3 - 2230

€/ GJ

27 - <<

14 - <<

~11 - <<

€/ GJ

24 - <<

5 - <<

~9 - <<

Environmental impact 1st gen bioenergy crop e

-

-/+

Environmental impacts herbaceous crop e +/+ a For the Netherlands, this only includes land currently in use as agricultural land and is based on the findings of de Wit and Faaij (2010). In Mozambique and Ukraine available land also includes other land comprising mostly grassland and shrubland (forest, protected areas, steep slopes, etc. are excluded) b This is ethanol from sugar beet in the Netherlands, from sugar cane in Mozambique and form wheat in Ukraine. c This is ethanol from Miscanthus in the Netherlands, torrefied wood pellets from eucalyptus in Mozambique, and Switchgrass ethanol in Ukraine. d The cost of ethanol production in Ukraine are the production cost of ethanol considering the lowest available feedstock cost derived from de Wit and Faaij (2010). The cost provided for Mozambique include transport from the conversion plant to the harbour, storage and long distance shipping, whereas these cost are excluded in the cost calculations of the ethanol production in the Netherlands and Ukraine. e The environmental impacts included in the Netherlands included GHG emission, and impact and soil water and biodiversity. In Ukraine only GHG emission were assessed. For Mozambique, the environmental impacts of bioenergy production were not assessed in this thesis.

IV.

What reliability can be obtained using the data available and the methods developed in this study?

In this thesis, a spatiotemporal analysis of potential developments in land availability; and the production potential, economic performance and environmental impacts of bioenergy production has been conducted. Spatial explicit and ex ante analysis comes with numerous uncertainties. Land use change drivers How, in which direction, and at what pace the key drivers of land use change will develop in different settings is uncertain. In addition, the direction, type, extent of correlation and the relative importance of drivers for the location of land use change (such as the agroecological suitability, accessibility, conversion elasticity, and neighbourhood) are generally 260

7.

Summary and conclusions

difficult to detect and may change over time; as well may the drivers themselves. Therefore, validation and calibration of the spatial detailed land use change model is crucial. Longitudinal assessment by monitoring both land use change drivers and land use patterns of different countries, in different stages of developments, with different biophysical and socio-economic characteristics would improve the understanding of the correlations between drivers and land use change and how these will develop. Cost developments and economic viability of bioenergy production In the long run, fossil fuels are expected to become more expensive which could contribute to the viability of bioenergy production. However, in recent years, the biomass feedstock production costs have been affected by increased costs of equipment, diesel, fertilizers and agrochemicals. These costs may continue to increase in line with historical global trends. In addition, production costs in countries such as Mozambique and Ukraine are low because of low cost of e.g. labour and land. These are expected to increase due to an increased pressure on land and due to economic development. On the other hand, it is expected that these cost increases could be counteracted by means of more efficient management (e.g. improved breeding, higher conversion efficiencies, etc.). Land use competition As bioenergy crop production was not included as a dynamic land use class, and a fixed order of allocation of the other land use classes was applied, the competition for land or indirect land use change was not modelled as such. However, in practice it is likely that bioenergy crop production will compete with other land use functions for best suitable areas. If competition between bioenergy crops and other land uses is to be modelled, the implementation of bioenergy crops should relate to a projected demand (e.g. national biofuel blending targets). In order to properly model competition between land uses, extensive information is required on market developments, price elasticities and policies. However, the key starting point in all land availability analyses was that competition for land is to be avoided, and therefore land was excluded if it is (potentially) in use for other land use functions. Data availability and quality The spatially explicit assessment of land use change, environmental impacts and economic performance requires large amounts of (spatial) data. The availability and quality of (spatial-) data has limitations and there are many inconsistencies between several statistical data sources, between spatial data sets and between statistical and spatial data. In addition, several spatial data sources have severe quality issues related to resolution, classification, consistency and documentation. Furthermore, for the analyses, data of several sources are combined, which are not necessarily consistent with each other. Examples are the data on agro-ecological suitability with the data on SOC, C:N ratio and 261

SHADES OF GREEN

climate data. Especially better spatial data on key inputs such as current land use, soil characteristics (e.g. sediment, SOC, C:N ratio, ground water table), the use and conditions of pastures, the agro-ecological suitability for several land uses classes would be required for improved accuracy of the results. Also, more consistency in the statistical data on land use, yield levels, agricultural production, livestock numbers, etc would contribute to improved reliability. Uniformity - heterogeneity In the analysis of the potentials, costs and environmental impacts are assessed assuming national uniformity in land use classes, (developments in) agricultural management and drivers for land use change. However, in practice considerable spatial heterogeneity in these parameters occurs. For example it is assumed that the land use class ‘cropland’ consist of a weighted summation of all crops produced, whereas in practice, the composition of crop rotations will be different for different areas and locations. It is assumed that the agricultural practices and the adoption of improved practices and related increases in efficiency are uniform for all different kind of producers and uniform for all agro-ecological suitability classes. Only the yield related management practices such as fertilizer input and harvesting are varied in relation to the agro-ecological suitability. However, there are currently large differences in agricultural practices and in spatial distribution of small and large scale produces. However, for the sake of the land use change modelling on the national scale, average agricultural practices and average increases in efficiency is assumed for the entire country. Impact relations The calculations of the environmental impacts are based on a broad range of input parameters. As all impacts are related to ecosystem functionality, they are heavily interlinked with each other. In some cases, impacts reinforce each other and in other cases, tradeoffs between impacts may occur. It is however complicated to prove causality and to quantify relations. Environmental impacts could be modelled more accurately when the modelling is more process based instead of making use of default values and factors and when better data is available. In addition, evaluating the significance of the caused impact is complicated, as both the dose-response relationship and the thresholds for actual damage and can be location, time and scale dependent. Analysis of uncertainties The approach demonstrated in this thesis provides possibilities to deal with uncertainties within the spatiotemporal modelling of potentials, cost and environmental impacts. The sensitivity analyses show to what extent uncertainties in input data results in changes in the results. The scenario approach enables the modelling of divergent assumptions of key drivers of land use change. The PLUC model can deal with uncertainties by enabling Monte 262

7.

Summary and conclusions

Carlo analysis on stochastic input parameters. The development in potential land use 2 change, costs and environmental impacts is calculated on a 1 km grid size level, but the above mentioned factors limit the accuracy of the results. However, the current available data are considered to have sufficient accuracy to distinguish patterns and hotspots of land use change, economic viability and economic performance. Therefore overall, the results can be used for a first screening to identify ‘go’ and ‘no-go’ areas. However, if the presented approach is to be used for the monitoring and certification of biomass production, better data and a finer modelling resolution are desired.

7.5 Recommendations for further research •









In order to assess the dynamics of land use competition, allocation of land use should also be based on relative economic performance of alternative land uses. In order to account for the dynamics in prices, supply and demand for biofuels, fossil fuels and agricultural commodities, land use change modelling should be interlinked with general and/or partial equilibrium models. Longitudinal assessment by monitoring both land use change drivers and land use patterns would improve the understanding of the correlation between drivers and land use change and allow for making more reliable projections of possible future land use patterns. More knowledge is required on the interrelations between different environmental impacts. In addition, the differentiation of dose-response relationships and thresholds for actual damage related to the biophysical context, time and scale requires further research. Impacts on water should ultimately be assessed on a water basin level to capture all relevant mechanisms that determine whether water is used unsustainably or not. Finally, more validation and coherence is required between the methods and indicators to quantify and monitor impacts on biodiversity of land use change on different spatial levels. Socio-economic impacts of bioenergy production are often strongly related to land use, economic performance and environmental impacts. These impacts should be assessed in an integrated way in order to identify the most suitable areas and means of production, and to be able to tackle more complex indicators, such as food security, in a quantitative manner. Land use, economic performance, and environmental and socio-economic impacts of bioenergy are directly related to the dynamics of the entire agricultural sector (including use of pasture lands and the livestock sector). Therefore, the effects of introducing crop production for bioenergy and the developments of other land use functions should be analysed in an integrated

263

SHADES OF GREEN





manner. In this thesis a first step is made by including the entire agricultural sector (including bioenergy crops) in the spatiotemporal assessment of the GHG emissions. This approach can also be applied for other impacts. For improved reliability of an ex ante assessment of ‘go’ and ‘no go’ areas of bioenergy production, better quality (spatial-) data in terms of resolution, accuracy and documentation is required. Especially data on current land use (including pastures and the intensity of their use), agro-ecological suitability, soil characteristics and climate are required. Biorefining of biomass for several applications such as food, feed, fibre, energy and chemicals offers opportunities for efficient use of biomass resources. Opportunities of innovative pathways need to be explored to assess the potential improvements of the overall efficiency and economic performance of biomass supply chains.

7.6 Market and policy recommendations •





In order to achieve the high potential deployment levels of biomass for energy and avoid iLUC, an increased production of bioenergy, feed and food needs to be balanced by improvements in agricultural management. In addition, the implementation of sustainability and policy frameworks is required to ensure good governance of land use and improvements in forestry, agricultural and livestock management. Large bioenergy potentials are found in less developed areas. In order to develop the large bioenergy potentials in these regions, a more integrated approach for sustainable development is required. This involves investments in the agricultural sector but also in education, infrastructure and development of markets. As developing countries are often characterised with a weak policy and institutional framework, bioenergy production in these countries brings increased responsibilities for other stakeholders (producers, users, certification bodies and governments) to ensure sustainable development. An ex ante analysis of the land availability, the economic viability and the environmental impacts contributes to the identification of go and no- go areas for bioenergy production. This enables a sound planning of land use, sustainable investment in bioenergy production capacity, and infrastructure over time. It could also help investors and policymakers to make realistic estimations of the economic viability of a project and it provides the ability to define the preconditions to comply with sustainability criteria. This could help to reduce investment risks and avoid large scale project failures.

264

7.





Summary and conclusions

Advanced certification of sustainable biomass with verifiable and quantifiable parameters requires improved tools to assess the environmental and economic performance and impacts on land use. The presented approach could contribute to an ex ante analysis of the sustainability of bioenergy production for certification, but better data and a finer modelling resolution are required. Ideally, policies on agriculture, environment, renewable energy and rural development need to be aligned in order to develop a sustainable bioenergy sector, which is one component of sustainable governance of land at large.

265

Sumário e Conclusões

Sumário e Conclusões

267

SHADES OF GREEN

7.1

Contexto da pesquisa

Todas as sociedades necessitam de serviços energéticos para suprir as necessidades humanas básicas e para facilitar processos produtivos (IPCC< 2011). Atualmente, 85% de toda a oferta de energia primária vêm de combustíveis fósseis (IEA 2010; IPCC2011). No entanto, sistemas energéticos predominantemente de origem fóssil não podem ser sustentados ao longo prazo por sua quantidade finita, distribuição desigual de recursos e sua grande contribuição para as emissões antropogênicas de Gases do Efeito Estufa (GEE). O uso de fontes de energia renováveis deverá ser crescente para substituir os combustíveis fósseis e tornar os sistemas de energia mais sustentáveis (IPCC, 2007). Aplicações modernas de biomassa contribuíram com 11 EJ em 2008, e é esperado que realizem um importante papel na oferta futura de energia (IEA e OECD 2011; IPCC 2011). Por exemplo, baseado em análise de cenários, o Painel Intergovenamental de Mudanças Climáticas (IPCC, em inglês) estima que a produção de 120-190 EJ em 2050 a partir de biomassa serão necessários para atingir as metas de mitigação de GEE relativos à estabilização da concentração de CO2 eq atmosférico a níveis menor que 440 ppm até 2100 (IPCC, 2011). No entanto, a crescente produção e utilização de bioenergia não é singularmente motivada pelo seu potencial de mitigação de GEE (se produzidos sustentavelmente), mas também pela fácil implementação de bioenergia em infraestruturas energéticas existentes, a versatilidade da biomassa como fonte, a diversificação de fontes de energia e subsequente segurança energética, potencial para contribuição para o desenvolvimento rural e o potencial para resgate de terras degradadas. Contudo, uma crescente implementação de plantações exclusivas para a produção de energia poderia ter impactos socioeconômicos e ambientais adversos como desmatamentos, perda de sumidouros de carbono, biodiversidade e outras funções e serviços ecossistêmicos, remoção de pessoas e a crescente competição por terra, água e outros fatores de produção, que podem levar a alta nos preços de alimentos. Muitos destes impactos estão relacionados com mudança do uso do solo (LUC, em inglês) (Wicke et al.; 2012). Para atingir o aumento desejado no uso de bioenergia, a competição entre alimentos e combustíveis deve ser evitada. Isto é possível balanceando a crescente produção de biomassa para energia com melhorias na gerência agrícola (Dornbur et al., 2010; Wicke et al., 2012). Ainda mais, as questões ambientais devem ser tratadas selecionando sistemas de bioenergia apropriados e aplicando planejamentos de uso da terra adequados (van Dam et al., 2010). A implementação de esquemas de sustentabilidade, por exemplo, certificação, poderia mitigar impactos ambientais negativos, e simultaneamente, contribuir para efetivação dos múltiplos objetivos do desenvolvimento sustentável. 268

Sumário e Conclusões

Em vários níveis, iniciativas para critérios de sustentabilidade, códigos de conduta e protocolos estão sendo desenvolvidos para tratar dos assuntos relativos à sustentabilidade. Atualmente, um problema tanto para o mercado quanto para governos estão direcionados para quais caminhos a serem tomadas visando atingir os critérios de sustentabilidade na prática e como os impactos podem ser quantificados de maneira verificável e confiável. Fortes melhorias nas análises de potencial espacial explícito e impactos são necessárias para permitir uma certificação efetiva e boa governança no uso da terra e do setor agrícola, tudo em relação direta com a crescente produção de biomassa e subsequente uso para energia e produção de materiais.

7.2

Objetivos e questões da pesquisa

Esta tese tem por objetivo examinar como potenciais custos e impactos ambientais da produção de bioenergia podem ser avaliados, evitando iLUC e levando em conta a variabilidade espaço-temporal do contexto bio-físico e sócio-econômico. Para este fim, as seguintes perguntas de pesquisa foram feitas: I. Como uma potencial disponibilidade de terra para plantações específicas para produção de energia pode ser avaliada, de forma explícita, espacial e temporalmente, uma vez que iLUC deve ser evitado e, portanto, levando em conta o desenvolvimento em outras funções de uso da terra? II. Como a viabilidade econômica e os impactos ambientais da produção de bioenergia podem ser avaliados espacialmente e temporalmente de maneira explicita, incluindo a localização da competição específica com outros usos da terra, o custo de produção de biomassa como matéria-prima e da logística da cadeia de custódia, e os impactos sobre as emissões de GEE, solo, água e biodiversidade. III. Quais são os potenciais, desempenhos econômicos e impactos ambientais da produção de bioenergia em diferentes contextos? IV. Que confiabilidade pode ser obtida utilizando os dados disponíveis e os métodos desenvolvidos neste estudo? As questões de pesquisa são tratadas nos Capítulos 2 a 6. Nos Capítulos 2 e 3, a viabilidade econômica e os potenciais impactos ambientais de cadeias produtivas regionais de bioenergia foram avaliados espacialmente, de maneira explícita, levando em conta a variação espacial na adequação agro-ecológica, no uso da terra atual e outros fatores biofísicos. As metodologias baseadas em GIS desenvolvidas nestes capítulos permitem avaliações espaciais explícitas, mas avaliações estáticas de custos e impactos ambientais. O norte da Holanda foi escolhido como uma área de estudo de caso por causa da alta competição por terra e e do seu uso intensivo e, por causa da disponibilidade adequada 269

SHADES OF GREEN

de dados espaciais detalhados. No Capítulo 4, um novo modelo de mudança no uso da terra (PLUC) foi desenvolvido para avaliar a disponibilidade de terras para as culturas de bioenergia, levando em conta fatores para a mudança de uso da terra específicos para o país. O modelo desenvolvido permite a avaliação espacial e temporal explícita de disponibilidade de terras para plantações para bioenergia e facilita avaliação espaçotemporal incerta com base em parâmetros de entrada estocásticos. No capítulo 5, o desenvolvimento em custos de matérias-primas de biomassa e a logística de cadeias de custódia de bioenergia foram analisados, levando em conta os desenvolvimentos no potencial de disponibilidade de terras e no aprendizado tecnológico. O acoplamento do modelo espaço-temporal do uso da terra, os custos espacialmente explícitos de produção de biomassa e logística e, o desenvolvimento de custos temporais, permitiram a avaliação espaço-temporal dos custos de fornecimento de bioenergia. No capítulo 6, o modelo de mudança de uso da terra desenvolvido foi adaptado para a Ucrânia e ampliado com um módulo de emissão de GEE, a fim de analisar os potenciais desenvolvimentos nas emissões de GEE de todo o setor agrícola, incluindo a implementação de culturas para produção de bioenergia e intensificação do setor agrícola. O acoplamento entre o módulo dinâmico de cálculo de emissões de GEE e o modelo de mudança de uso da terra permitiram a avaliação de impacto espaço-temporal de GEE da mudança total de uso da terra. Moçambique e Ucrânia foram escolhidos como áreas de estudo de caso devido ao seu alto potencial de produção de bioenergia relacionado com a baixa densidade populacional e do clima favorável para a produção de biomassa e, também devido à diversidade das condições ambientais e sócio-econômicas dos países. Nos Capítulos 2, 3, 5 e 6, o desempenho das culturas típicas para a produção de biocombustíveis de primeira e segunda geração é analisado. Na maioria dos os estudos de caso, assume-se essas culturas são usadas para a produção de etanol a fim de permitir a comparação dos potenciais, custos, e dos impactos ambientais para opções comuns de biocombustíveis de primeira e segunda geração. A complexidade e o nível de integração das metodologias desenvolvidas nesta tese aumentam de capítulo a capítulo e evolui a partir de modelagem espacialmente explícita e modelagem estática (no Capítulo 2, desempenho econômico e no Capítulo 3, impactos ambientais) para modelagem espaço-temporal e modelagem dinâmica (no Capítulo 4, mudança de uso da terra e no Capítulo 5 desenvolvimento de fornecimento de custo). No capítulo 6, a modelagem espaço-temporal de uso da terra é integrada com a modelagem dinâmica de emissão de GEE, que permite a avaliação integrada de impacto espaçotemporal. Em todos os capítulos, as limitações dos métodos e dos dados utilizados nas avaliações são discutidas. A Tabela 7.1 apresenta uma visão geral dos capítulos e as questões de pesquisa abordadas.

270

Sumário e Conclusões

Tabela 7.1: Visão geral das configurações dos capítulos da tese e as questões de pesquisa tratadas neles.. Capítulos 2 3 4 5 6

7.3

Questões da pesquisa I II III

Potencial distribuição espacial e desempenho econômico de cadeias regionais de biomassa Variação espacial de impactos ambientais em cadeias regionais de biomassa Modelagem espaço-temporal de uso da terra para avaliar plantações para produção de energia Curvas de custo-fornecimento espaço-temporal para produção de bioenergia Análise espaço-temporal integrada da mudança do uso da terra para agricultura, potencial de produção de bioenergia e balanços de GEE relacionados

































IV

Resumo dos resultados

No Capítulo 2 são abordadas as questões de pesquisa II, III e IV, através da análise da variação espacial da viabilidade econômica para a produção de etanol a partir de miscanto e beterraba-sacarina no norte da Países Baixos. A competitividade dos cultivos bioenergéticos foi avaliada através da comparação do Valor Presente Líquido (NPV, Net Present Value em inglês) de culturas perenes, rotações convencionais e sistemas de rotação que incluem anos adicionais de beterraba-sacarina, seguido duma comparação do custo de produção de bioetanol com o preço médio da gasolina. O uso actual do solo e a sua aptidão para cultivo convencionais e energéticos foram mapeados através dum Sistema de Informação Geográfica (SIG), a partir do qual foi possível determinar as zonas onde são mais prováveis a ocorrência de alterações do uso do solo, de acordo com a distribuição espacial de rentabilidade econômica. Os custos de produção de bioetanol incluem os custos associados à lavoura, colheita, transporte e conversão em etanol. O NPV e o custo de produção de matéria-prima foram calculados para sete classes de aptidão do solo. Os resultados mostram uma grande variação espacial tanto do custo de produção da matéria-prima vegetal como da rentabilidade da sua produção em comparação com os cultivos agrícolas convencionais. De acordo com os preços atuais de mercado, as culturas energéticas não são competitivas com os sistemas convencionais de cultivo em solos classificados como "aptos". Em solos menos adequados, o retorno das culturas geridas intensivamente é baixo e consequentemente as culturas perenes obtêm maiores NPVs do que as rotações convencionais. Os resultados mostram ainda que os custos mínimos de produção de matérias-prima vegetal são de 5,4 €/GJ para miscanto e 9,7 €/GJ para a beterraba-sacarina, dependendo da aptidão do solo. Os custos do etanol produzido a partir de miscanto (24 €/GJ) são mais baixos que o etanol de beterraba açucareira (27 €/GJ), mas ainda assim o bioetanol elaborado a partir de cultivos produzidos domesticamente não é competitivo com a gasolina (12,3 €/GJ), mediante as circunstâncias 271

SHADES OF GREEN

actuais. No entanto, caso os preços do petróleo aumentem ou a aprendizagem tecnológica e biorefinação resultarem em cadeias de fornecimento mais eficazes em termos de custos, a competitividade poderá melhorar. Neste capítulo é apresentada uma metodologia genérica para identificar locais atractivos de um ponto de vista econômico para a produção de cultivos bioenergéticos, tendo em conta a variação espacial do uso actual do solo e fatores biofísicos, tais como características do solo e disponibilidade de água. No Capítulo 3 são abordadas as questões de pesquisa II, III e IV. Neste capítulo, a variação espacial dos potenciais impactos ambientais resultantes da produção de cultivos de bioenergéticos é avaliada quantitativamente. O cultivo de beterraba-sacarina e miscanto para produção de bioetanol no norte da Países Baixos é utilizado como caso de estudo. Os impactos ambientais incluem: emissões de gases de efeito estufa (GEE) ,durante o ciclo de vida e resultantes da alteração directa do uso do solo, qualidade do solo, quantidade e qualidade da água e biodiversidade. Para cada impacto, foram selecionados indicadores e métodos adequados com base numa revisão bibliográfica extensiva, os quais foram adaptados sempre que necessário de forma a poderem ser aplicados de forma espacialmente explícita. A variação espacial dos impactos ambientais resultante da heterogeneidade espacial do contexto físico é avaliada através de um Sistema de Informação Geográfica (SIG). Este caso de estudo mostra que existem grandes variações espaciais nos impactos ambientais resultantes da introdução de cultivos bioenergéticos. De forma geral, a beterraba-sacarina é o cultivo que causa relativamente mais impactos ambientais negativos, especialmente em áreas de pastagem. Nessas áreas, as emissões de GEE resultantes das alterações directas do uso do solo podem ir até 148 kg CO2-eq GJetanol 1 -1 e o risco de erosão do solo até 9 t ha . Além disso, existe um alto risco de perda de biodiversidade nestas áreas. Em termos de impactos positivos, observa-se um decréscimo -1 de 75 mg l na concentração de NO3 nas águas subterrâneas e uma diminuição de 100 mm no défice sazonal de água. Quando as terras agrícolas são convertidas para o cultivo -1 de miscanto, regista-se uma redução das emissões de gases de -159 kg CO2 GJetanol , uma -1 redução do risco de erosão do solo de 4 t ha , e uma redução da concentração de NO3 de -1 53 mg l , para além de efeitos positivos para a biodiversidade. Mas por outro lado, a depleção sazonal de água pode aumentar até 150 mm. Para ambas as culturas, as zonas de pastagens úmidas na parte ocidental da área de estudo surge como a zona de maior incidência de impactos negativos. Através da combinação espacial da avaliação dos impactos ambientais, é possível concluir que existem um compromisso de vantagens e desvantagens entre os vários impactos ambientais: em todas as áreas ocorre pelo menos um dos impactos ambientais negativos. Este estudo apresenta um framework para identificar as áreas com potenciais impactos ambientais negativos da produção de cultivos

272

Sumário e Conclusões

bioenergéticos e áreas onde a produção destes cultivos têm impactos ambientais limitados ou mesmo positivos. No Capítulo 4 são abordadas as questões de pesquisa I, III e IV, sendo desenvolvido um modelo para avaliar de forma espacialmente explícita os futuros desenvolvimentos na disponibilidade de terras para culturas bioenergéticas, tendo em conta a prevenção da ocorrência de competição entre funções de uso do solo, de forma a evitar a alterações indiretas do uso do solo. Especificamente para cada país, são identificados e incluídos no modelo os fatores que influenciam e determinam o uso do solo relativamente à outras funções, tais como produção de gado, alimentos e matérias-primas, assim como a incerteza no futuro desenvolvimento desses fatores. Um modelo espaço-temporal de uso do solo baseado em PCRaster é demonstrado através de estudo de caso sobre a evolução da disponibilidade de terras para cultivos bioenergéticos em Moçambique entre 20052030. A evolução dos principais fatores para uso agrícola do solo (procura de produtos alimentícios, produtos de origem animal e matérias-primas) foi avaliada com base nos desenvolvimentos previstos em termos de população, dieta, PIB e taxa de autosuficiência. Dois cenários foram desenvolvidos: um cenário business-as-usual e um cenário progressivo. A alocação de terras foi baseada num conjunto de fatores de amplitude específico para cada classe de uso do solo. A dinâmica de alteração de uso do solo foi 2 mapeada em uma grelha de 1 km para cada ano até 2030. 7,7 Mha no cenário BAU e 16,4 milhões de hectares no cenário progressivo podem ficar disponíveis para a produção de bioenergia em 2030. Com base na análise de Monte Carlo, foi encontrado um intervalo de confiança de 95% em termos da quantidade de terras disponíveis e da probabilidade espacialmente explícita de disponibilidade de terra. A abordagem bottom-up, a inclusão de vários usos dinâmicos do solo e diversos fatores de aptidão, e a possibilidade de simular incerteza permite melhorar significativamente a modelação integrada da disponibilidade de terras para produção de bioenergia. Além disso, permite identificar oportunidades e explorar as condições previas para a produção em larga escala de bioenergia, evitando alterações indiretas do uso do solo. No Capítulo 5 são abordadas as questões de pesquisa II, III e IV, sendo avaliada de uma forma espacialmente explícita a evolução do custo e do potencial de produção de bioenergia ao longo do tempo. Esta avaliação é baseada nos desenvolvimentos de disponibilidade de terras, na aptidão do solo que se encontra ou que pode vir a ficar disponível, no custo desagregado de produção de culturas bioenergéticas, na distância de transporte da matéria-prima para a planta de conversão, no custo de conversão, na distância de transporte da planta até ao porto de exportação e nos custos de transporte marítimo internacional. As cadeias de fornecimento de pelotas (torrefeitas) de eucalipto e de etanol feito a partir de cana-de-açúcar em Moçambique são aplicados como caso de 273

SHADES OF GREEN

estudo. Os desenvolvimentos na disponibilidade de terras para culturas energéticas em Moçambique são baseadas nos resultados obtidos no Capítulo 4. Observa-se uma grande variação espacial nos custos da cadeia de fornecimento, como consequência da variação espacial dos custos de produção de matéria-prima e dos custos de transporte primário e secundário. As zonas mais promissoras para produção de eucalipto e cana estão localizadas no centro-sul, centro, e nordeste de Moçambique, onde as condições agroecológicas são relativamente mais favoráveis, suficiente matéria-prima pode ser produzido de forma a atender os requisitos duma planta de conversão, e onde se encontra infra-estrutura disponível. No cenário progressivo, o potencial total calculado para a produção de pelotas de eucalipto equivale a 3200 PJ em 2030, dos quais 2500 PJ poderiam ser exportados para a Europa por um preço de mercado inferior a 8 €/GJ. Para a produção de etanol de cana, foi determinado um potencial de 850 PJ, do qual 500 PJ poderiam ser exportados por um preço de mercado inferior a 30 €/GJ. A localização da produção é o factor-chave para uma produção rentável. Isto é especialmente verdadeiro em países com uma grande heterogeneidade em termos de aptidão agro-ecológica e a com uma baixa disponibilidade de infra-estrutura. Caso a rede de estradas e infra-estrutura ferroviária fosse melhorada, o custo de logística poderia ser reduzido e mais áreas poderiam tornarse economicamente viáveis para produção de bioenergia. Este estudo demonstra uma metodologia que permite a avaliação do desenvolvimento de potenciais bioenergéticos e dos seus custos de produção ao longo do tempo de uma forma espacialmente explícita. Como os impactos ambientais e sócio-económicos das cadeias de fornecimento de bioenergia são altamente relacionados com o contexto biofísico e sócio-económico do local de produção, o estudo espacialmente explícito do potencial de produção de bioenergia é uma abordagem adequada para conceber, optimizar e analisar os impactos e sustentabilidade das cadeias de fornecimento de bioenergia. No Capítulo 6 são abordadas as questões de pesquisa I, II, III e IV através da análise espacialmente explícita dos futuros desenvolvimentos do potencial bioenergético e de redução da emissão de gases de efeito estufa (GEE) na Ucrânia no período de 2010-2030, tendo em conta o desenvolvimento e emissões de outras funções agrícolas de uso do solo. O desenvolvimento em termos de requisitos de terra para a produção de alimentos e rações é analisada espacialmente numa base anual através do modelo de alterações de uso do solo PCRaster (PLUC, PCRaster Land Use Change em inglês). O modelo foi adaptado para a Ucrânia, através do ajuste das classes dinâmicas de uso do solo, dos fatores de aptidão e suas características de acordo com a situação ucraniana. Dois cenários foram avaliados para o período de 2010-2030: um cenário business-as-usual, em que se assume uma continuidade das tendências actuais de produtividade e um cenário progressivo, no qual se projecta uma convergência dos níveis de rendimento da Ucrânia com os países da Europa Ocidental. No cenário progressivo, 32,2 Mha de terra poderiam tornar-se 274

Sumário e Conclusões

disponíveis para a produção de cultivos energéticos em 2030. Estas projecções acerca do uso do solo serviram posteriormente para determinar o balanço espacio-temporal de GEE, incluindo emissões de CO2, N2O e CH4 relacionadas com as alterações na gestão e uso do solo, bem como a redução de emissões de GEE através da substituição de combustíveis fósseis por produção de bioetanol a partir de trigo e switchgrass (Panicum virgatum). O módulo espaço-temporal de GEE produz mapas espacialmente explícitos (resolução de 1 2 km ) das emissões de cada um dos GEE numa base anual. São obtidos níveis elevados de sequestro de carbono com cultivos de switchgrass ou quando a vegetação natural é regenerada nas terras agrícolas abandonadas. As áreas abandonadas estão localizadas principalmente no norte, leste e sul da Ucrânia. As emissões totais de N2O aumentam quando a terra agrícola é cultivada com switchgrass ou trigo, especialmente em áreas onde as terras agrícolas se expandem à custa de áreas mistas de pastagem e terras agrícolas. Os resultados mostram que um balanço total acumulado de GEE de -0,8 GT CO2eq para o trigo e -3,9 GT CO2-eq para switchgrass poderia ser alcançado em 2030 no cenário progressivo. O balanço negativo do trigo é também causado pela re-crescimento de vegetação natural em áreas agrícolas abandonadas que são excluídas para a produção de cultivos de bioenergéticos. Com a implementação de medidas adicionais de redução de GEE, nomeadamente para reduzir as emissões de CO2 e N2O agrícola (por exemplo, lavoura reduzida, aumento da adição de carbono orgânico e melhor utilização de fertilizantes), o balanço cumulativo de GEE poderia mesmo aumentar até -2,6 GT CO2-eq para o trigo e -5,1 GT CO2-eq para switchgrass em 2030. Quando a terra disponível é utilizada para o re-crescimento de vegetação natural, uma quantidade considerável de carbono será acumulada sob a forma de biomassa e de carbono orgânico, podendo chegar a -4,4 GT CO2-eq em 2030. No entanto, esta fixação do carbono só poderá ser obtida em detrimento de culturas energéticas. Considerando o cenário progressivo e um período até 2100, a redução total acumulada de GEE é estimada em -5,5 ± GT CO2-eq por recrescimento de vegetação natural, -6 GT CO2-eq para o trigo e -15 GT CO2-eq para switchgrass. O módulo espacio-temporal de GEE em conjunto com o modelo PLUC permite a modelação espacialmente explícita e dinâmica das emissões de GEE resultantes de gestão e alterações no uso da terra relacionadas com a implementação da produção de culturas de bioenergia. De acordo com a metodologia desenvolvida e as conclusões apresentadas nos capítulos 26, serão dadas respostas às principais questões de pesquisa, assim como recomendações para políticas e futura investigação.

275

SHADES OF GREEN

7.4

Principais resultados e conclusões

I. Como pode o desenvolvimento potencial da disponibilidade de terra para cultivos energéticos ser avaliada de forma espacialmente e temporalmente explícita, tendo em conta o desenvolvimento em outras funções de uso do solo e que deverão ser evitadas alterações indiretas do solo (iLUC em inglês)? Neste estudo, um novo modelo de alteração de uso do solo (PLUC) foi desenvolvido para avaliar o desenvolvimento da disponibilidade de terras para cultivos bioenergéticos a um 2 nível espacial detalhado (1 km ), tendo em conta a dinâmica e as incertezas dos principais fatores de mudança de uso do solo. Os principais fatores para o desenvolvimento da demanda de produtos agrícolas numa região são os desenvolvimentos na população, mensurados através do PIB, consumo de alimentos e utilização de materiais per capita e taxa de auto-suficiência. A quantidade de terra necessária para atender à demanda total depende da eficiência do setor agrícola e aptidão agro-ecológica das áreas de produção. Uma vez que é incerta a forma como os fatores de alteração do uso do solo (LUC factors em inglês) vão evoluir, o uso de cenários poderá ser uma abordagem adequada para explorar potenciais desenvolvimentos a longo prazo nos fatores LUC. A quantidade total de terra necessária para atender à demanda de madeira, alimentos e produtos animais está directamente relacionada com a produtividade da localização para cada classe específica de uso do solo (por exemplo, terras agrícolas, pastagens, floresta). A alocação de terra para usos do solo dinamicos baseia-se na aptidao do local para cada classe específica de uso do solo, a qual se encontra definida por uma combinação de fatores de aptidão espacialmente explícitos, nomeadamente a aptidão agro-ecológica, acessibilidade, elasticidade de conversão e relações vizinhança. Para cada factor de aptidão, a direcção, o tipo e a extensão da correlação necessitam ser determinados. Os fatores de aptidão, suas características e sua importância relativa são específicos para cada região e classe de uso do solo. As áreas que não são adequadas (por exemplo, encostas íngremes) ou para as quais a conversão em terras agrícolas não é permitida (áreas de conservação, por exemplo) são excluídas. O ponto de partida de alocação de uso da terra é o mapa actual de uso do solo, calibrado de acordo com as estatísticas sobre a demanda de alimentos para consumo humano, rações e matérias-primas, produtividade agrícola e a distribuição das classes de uso da solo em termos de aptidão agroecológica. A simulação da alocação de terra para as classe de uso do solo é feita em passos de tempo de um ano. A alocação total para um passo de tempo é dada como concluída quando todas as classes de uso do solo tiverem sido alocadas e a produção dessas classes de uso do solo forneça a demanda total para aquele ano específico. A modelação inclui um ciclo de feedback: o mapa de uso do solo resultante da 276

Sumário e Conclusões

alocação no passo de tempo t serve como ponto de partida para a alocação no passo de tempo t+1. O resultado deste tipo de modelação dinâmica das alterações do uso do solo 2 são mapas representado o uso do solo a um nível espacial detalhado (1km ) e a disponibilidade de terras para cultivos bioenergéticos para cada ano, durante o período de modelação. A principal vantagem deste framework de modelação é a sua capacidade de lidar com dados estocásticos. Isto permite a execução de análise espacio-temporal Monte Carlo (MC), a qual avalia a propagação da incerteza. O modelo PLUC pode modelar estocasticamente séries temporais (por exemplo, a demanda e a produtividade dos cultivos agrícola), parâmetros de entrada espacial (densidade populacional e produtividade), e as características dos fatores de aptidão (por exemplo, a distância máxima de efeito de distância, no caso do factor distância às estradas). A modelação estocástica permite análises de sensibilidade dos resultados às incertezas nos parâmetroschave. Os resultados são apresentados em mapas representando a probabilidade de 2 disponibilidade de terra num local específico (célula de 1km ) num passo de tempo específico (ano). O modelo de uso do solo desenvolvido neste estudo é uma ferramenta avançada para avaliar a dinâmica futura do uso do solo e a disponibilidade de terras para cultivos energéticos. A aplicação de cenários sobre os principais fatores de mudança no uso do solo e o paradigma “alimentos primeiro” permitem uma avaliação dos potenciais de biomassa que podem ser alcançados sem a ocorrência de competição com a produção de alimentos e rações para animais, assim como as condições necessárias para materializar esses potenciais. A abordagem bottom-up, o número de usos dinâmicos do solo, o portefólio diversificado de fatores LUC e fatores de aptidão e a possibilidade de modelar incerteza é um passo em frente na modelação da disponibilidade de terras para cultivos energéticos. Como os rendimentos de biomassa, os custos de produção, logística, e os impactos ambientais estão fortemente relacionados com as características especificas de cada localização, tais como condições biofísicas (aptidão agro-ecológica, disponibilidade de infra-estrutura, propriedades do solo, condições climáticas, etc), uma avaliação espacialmente explícita da disponibilidade de terras para cultivos bioenergéticos é um pré-condição importante para projectar cadeias e logística de abastecimento de bioenergia, determinar o potencial de produção de bioenergia e impactos ambientais e socio-económicos. O modelo foi adaptado e demonstrado para Moçambique e Ucrânia. Contudo, é um modelo flexível que pode ser usado para outros países ou regiões, caso os dados, regras de alocação e características dos fatores de aptidão sejam adaptados. 277

SHADES OF GREEN

II. Como podem a viabilidade econômica em termos de custos de produção e vantagens competitivas, e impactos ambientais tais como impactos nas emissões de GEE, solo, água e biodiversidade, da produção de bioenergia, ser avaliados de forma espacialmente e temporalmente explícita? Na fase inicial do estudo, foram desenvolvidos métodos para avaliar a performance econômica e ambiental duma forma espacialmente explícita, tendo em conta a variabilidade espacial do contexto biofisico. Numa fase posterior, um modelo espaciotemporal foi desenvolvido para avaliar a viabilidade econômica e as emissões de GEE de forma a levar em conta tanto a variabilidade espacial como a temporal. A viabilidade econômica da produção de bioenergía depende das vantagens competitivas da produção de cultivos bioenergéticos em comparação com outro usos do solo e a competitividade da produção de bioenergia em comparação com um sistema de energia de referência. Relativamente aos custos, três fatores-chave determinam o custo da produção de bioenergia: custo da produção de matéria-prima vegetal, custos de logística da cadeia de fornecimento e o custo e eficiência da tecnologia de conversão. Os custos da produção de matéria prima podem ser avaliados através do calculo do valor presente liquido (NPV em inglês) de todos os componentes dos custos (terra, emprego, maquinaria, consumo de recursos) e o rendimento da produção de biomassa durante o tempo de vida da plantação de biomassa. A variação espacial do rendimento e custos associados assim como a competição com o uso do solo de referência podem ser calculados através da combinação do mapa de uso do solo e mapas de aptidão agroecológica específicos para cada cultivo. A variação espacial no custo de logística da biomassa pode ser analisado através da informação acerca da escala de conversão da planta e a informação espacial relativamente á disponibilidade de terra, níveis de rendimento, custos de operação e manutenção (O&M) e custos de consumo de energia. O custo de produção de biomassa e de conversão sofrem alterações ao longo do tempo, devido a aprendizagem tecnológica. A integração de projeções de aprendizagem tecnológica no calculo dos custos, interligada com a modelação espacio-temporal do uso do solo permite uma avaliação espacialmente e temporalmente explícita da performance econômica das cadeias de fornecimento de bioenergia. A seleção dos impactos ambientais da produção de bioenergia analisados nesta tese foi baseada nas áreas de preocupação identificadas em varias iniciativas nacionais e internacionais relativamente a criterios de sustentabilidade da produção de bioenergia (EC 2009; NEN 2009; RSB 2010). Os impactos ambientais tomados em consideração foram as emissões de GEE (durante o ciclo de vida e resultantes de alterações do uso do solo) e 278

Sumário e Conclusões

impactos no solo, água e biodiversidade. Os impactos ambientais foram determinados quantitativamente de uma forma espacialmente explícita através do implementação/adaptação de metodologias existentes para uma análise espacial detalhada. Os impactos na água foram avaliados através do uso dum balanço de água simples e tendo em conta informação espacial detalhada relativamente a alterações do uso do solo, precipitação efectiva e potencial de evapo-transpiração. Os impactos no solo foram avaliados através do uso da Equação de Erosão pelo Vento (WEQ, Wind Erosion Equation em inglês) e tendo em conta a variação espacial nas alterações do uso do solo, características da vegetação, do solo, vento, precipitação e temperatura. Os impactos na biodiversidade foram explorados através das alterações dos indicadores de Abundância Media de Espécies (MSA, Mean Species Abundance em inglês) e de Elevado Valor Natural (HNV, High Nature Value em inglês) resultantes do uso e gestão do solo e da distribuição de áreas naturais e espécies em perigo. Numa fase inicial, o modelo Miterra (Lesschen 2008; Velthof et al. 2009) foi usado para avaliar as emissões de GEE e o impacto na qualidade da água resultantes das alterações no uso do solo. Este é um modelo deterministico, estático e baseado em vectores que simula o balanço de azoto e fosforo, emissões de NH3, N20, NOx e CH4, lixiviação de N, concentração de NO3 nas águas subterrâneas e alterações no sequestro de carbono pelo solo e biomassa devido a alterações na gestão e uso do solo. Numa fase posterior, um novo modelo foi desenvolvido com o objectivo de avaliar as emissões de GEE de forma 2 dinâmica a um nível espacial detalhado (1 km ). O cálculo do ciclo de azoto foi baseado principalmente nos métodos desenvolvidos por Lesschen (2008) e Velthof et al. (2009), ainda que adaptados para o cálculo em raster. Para além disso, em vez de ser usado o factor de emissão de referência proposto pelo IPCC, foram implementados os fatores de emissão de N2O desenvolvidos por Lesschen et al. (2011), os quais são específicos para a fonte de azoto, tipo de solo e uso do solo. Os fatores de lixiviamento e escoamento foram também ligeiramente adaptados. Os cálculos da emissões de carbono são baseados no método proposto por IPCC (2006). Através da integração do módulo de emissões de GEE com o modelo espacio-temporal de uso do solo, as emissões de GEE podem ser calculadas de forma espacialmente e temporalmente explícita. Este novo modelo é passível de levar em conta a variação espacial de uso do solo, níveis de rendimento, características do solo, clima e declive; assim como a variação temporal em termos de uso do solo, gestão e níveis de rendimento. Para além disso, permite identificar as melhores áreas para intensificação de agricultura e as áreas mais aptas para a produção de cultivos bioenergéticos. Acima de tudo, pode contribuir para avaliar sob que condições o balanço de GEE poderia ser optimizado em conjunção com a salvaguarda da produção suficiente de alimentos e rações.

279

SHADES OF GREEN

Os métodos e modelos desenvolvidos fornecem uma abordagem para a identificação ex ante das áreas onde a implementação da produção de bioenergia é ou poderia vir a tornar-se economicamente viável e em que áreas poderia ter um impacto negativo pequeno ou até mesmo um impacto ambiental positivo. Através da comparação entre a variação espacial de vários impactos ambientais e da viabilidade econômica é possível uma avaliação do compromisso entre impactos ambientais e entre o desempenho ambiental e econômico. A integração destes modelos e métodos permite a identificação de áreas onde a produção de bioenergia deve ou não deve ser permitida, dum ponto de vista económico e ambiental, e em que a competição pela terra é evitada. III. Quais são os potenciais, desempenho econômico e impactos ambientais da produção de bioenergia em diferentes condições? Caso se pretenda evitar as alterações indiretas de uso do solo (iLUC em inglês), a quantidade de bioenergia que pode ser produzida depende dos desenvolvimentos da demanda por outros produtos agrícolas, da taxa de intensificação do setor agrícola e da aptidão do solo que se encontra disponível para produção de cultivos energéticos. Nos Países Baixos, a população encontra-se com uma taxa de crescimento baixa e presume-se que o consumo de alimentos estabilize nos níveis actuais. Se for assumido que a taxa de auto-suficiência se mantenha estável, pode-se esperar que a demanda total por produtos alimentares e rações aumente apenas ligeiramente ao longo do tempo. Os Países Baixos tem um dos setores agrícolas mais eficientes e tecnologicamente avançados da Europa (de Wit et al. 2011). Logo, as diferenças de rendimento e as oportunidades derivadas da melhoria da eficácia são relativamente limitadas (quando comparado com Ucrânia ou Moçambique). Isto resulta numa disponibilidade de terra para cultivos bioenergéticos relativamente baixa, entre 41 a 51 kha em 2030 (de Wit and Faaij, 2010; Fischer et al. 2010a; Fischer et al. 2010b). Tendo em conta a média regional dos níveis de rendimento de miscanto e beterraba-sacarina, 6 a 9 PJ de etanol poderiam ser produzidos em 2030. Uma vez que não foi determinado em que locais a terra poderia vir a tornar-se disponível, não foi possível calcular nenhuma estimação precisa do potencial de produção de cultivos bioenergéticos. Os custos de produção de etanol, baseado no custo mínimo de produção de matéria-prima, são de 24€/GJ para miscanto e 27€/GJ para beterrabasacarina (considerando a performance das tecnologias disponíveis no curto prazo). Em Moçambique, a demanda por produtos agrícolas irá aumentar devido a um grande aumento da população e melhorias na dieta. Por outro lado, existe um grande potencial para melhorar a produtividade agrícola, uma vez que esta é atualmente muito baixa. Quando um cenário business-as-usual (BAU) é assumido, no qual as tendências actuais de produtividade agrícola se mantêm, 7.7 Mha poderiam ficar disponíveis; num cenário progessivo, em que a produtividade agrícola aumenta consideravelmente, 16.4 Mha 280

Sumário e Conclusões

poderiam ficar disponíveis para a produção de bioenergia em 2030. Quando toda a terra disponível é usada para produção de eucalipto, 1340 PJ de pelotas torrefeitas poderiam ser produzidas num cenário BAU e 3200 PJ num cenário progressivo, em 2030. Quando a terra disponível é usada para cultivar cana-de-açúcar para a produção de etanol, 350 PJetanol no cenário BAU e 850 PJetanol no cenario progessivo poderiam ser produzidos em 2030. O custo mínimo de produção de pelotas torrefeitas é de 5 €/GJ e para o etanol de cana-de-açúcar de 14 €/GJ. As áreas que apresentam custos mínimos são aquelas onde a produtividade é maior, onde suficiente matéria-prima pode ser produzida de forma a ir de encontro aos requisitos da planta de conversão e que se encontram próximas de infraestrutura e portos. Se a rede de estradas e caminhos de ferro for melhorada, o custo de logística pode ser ainda reduzido e mais áreas podem vir a tornar-se economicamente viáveis para a produção de bioenergia. Na Ucrânia, espera-se que a demanda de produtos agrícola venha a aumentar ligeiramente. O tamanho da população encontra-se a decrescer e o consumo de alimentos tem tendência para estabilizar, mas os níveis de exportação deverão aumentar. Apesar das favoráveis condições agro-ecológicas, a produtividade do setor agrícola é baixa, o que oferece um potencial elevado de melhoria do rendimento. No cenário BAU, no qual ocorrem pequenas melhorias, a área disponível para produção de bioenergia é limitada (0.03 Mha). No entanto, no cenário progressivo 32.2 Mha poderiam ficar disponíveis para produção de cultivos energéticos em 2030. Caso esta área fosse usada para produção de etanol a partir de switchgrass, 3 PJetanol no cenário BAU e 2230 PJetanol no cenário progressivo poderiam ser produzidos em 2030. Se o trigo for cultivado na área abandonada, 1 a 2370 PJetanol poderiam ser produzidos em 2030. Com base no custo mínimo de produção de matéria-prima providenciado por Wit e Faaij (2010), os custos de produção de etanol variam entre 9 e 11 €/GJ. No cenário progessivo, foi determinado um grande potencial de terra disponível e de produção de bioenergia na Ucrânia. No entanto, no cenário BAU uma área muito pequena encontra-se, e ficará, disponível para cultivos energéticos. Uma vez que a Ucrânia é constituída predominantemente por terra agrícola (79%) e a maioria da área restante consiste em florestas (15%) ou usos do solo estáticos (5%, por exemplo: áreas de conservação, zonas construídas), a área disponível para produção de cultivos energéticos émuito baixa, a não ser que a quantidade de terra necessária para produção de alimentos e gado aumente, através dum aumento da eficiência. Em Moçambique, existe atualmente uma disponibilidade de terra relativamente grande (9 Mha). Para além do uso actual de terra para agricultura (20%), floresta (60%) e usos estáticos (9%), aproximadamente 10% da área total poderia ser usada para cultivos bioenergéticos. No entanto, a disponibilidade de terra para cultivos bioenergéticos e zonas florestais diminui quando a 281

SHADES OF GREEN

quantidade de terra necessária para produção de alimentos e rações se expande devido a um aumento da demanda, e uma baixa a inexistente melhoria na eficiência agrícola é alcançada. Devido a grandes diferenças em termos de características biofísicas, condições socioeconômicas, desenvolvimentos históricos, nível de desenvolvimento e contexto politico especifico, a dinâmica de alteração do uso do solo em Moçambique e Ucrânia são determinados por diferentes fatores, ou alguns fatores tem um papel mais ou menos importante. Dada a a grande dependência na agricultura de subsistência, a falta de infrastrutura e a dependência em mercados locais, espera-se que a terra agrícola em Moçambique se concentre em áreas mais densamente povoadas, próximas das principais cidades e da rede de estradas. Na Ucrânia, o número de pessoas que depende directamente da agricultura é menor, existe uma rede de estradas relativamente densa e encontra-se atualmente em curso uma mudança para unidades de produção maiores e mais comerciais. Assim, a aptidão agro-ecológica é aqui um factor-chave para a localização das terras agrícolas e é observada uma menor correlação entre densidade populacional ou acessibilidade. Devido a estas diferenças, os parâmetros no modelo PLUC (fatores, regras de alocação, fatores de aptidão, suas características e relativa importância) necessitam ser adaptados para cada situação. A avaliação do desenvolvimento dos custo das cadeias de fornecimento mostra que os custos podem ser significativamente reduzidos através da melhoria dos rendimentos, da logística e das tecnologias de pré-tratamento e conversão. A avaliação integrada dos impactos ambientais no norte dos Países Baixos mostra uma grande variação espacial nos impactos. Não existem áreas onde apenas efeitos positivos ocorram quando os cultivos bioenergéticos são introduzidos e existe um compromisso de vantagens e desvantagens entre impactos. Este caso de estudo demonstra que existem grandes variações espaciais nos impactos ambientais resultantes da introdução de cultivos bioenergéticos. No geral, a beterraba-sacarina provoca relativamente mais impactos negativos, especialmente em zonas de pastagem. Nestas áreas, as emissões de GEE resultantes de alterações no uso do solo podem ir até 148 kg/GJetanol e o risco de erosão do solo pode aumentar até 9 ton/ha. Para além disso, há um maior risco de perda de biodiversidade nestas áreas. Como impacto positivo, nestas zonas regista-se uma potencial diminuição de 75 mg/l na concentração de NO3 nas águas subterrâneas e uma possível diminuição de 100mm no deficit sazonal de água. Quando as terras agrícolas são convertidas para a produção de miscanto, observa-se uma redução das emissões de GEE de -159 kg/GJetanol, uma diminuição do risco de erosão do solo de 4 ton/ha, uma potencial redução da concentração de NO3 de 53 mg/l, assim como um efeito positivo na 282

Sumário e Conclusões

biodiversidade mas com a possibilidade de aumento da depleção de água em 150 mm. Os padrões espaciais dos impactos estão maioritariamente relacionados com o padrão espacial de uso actual do solo; por exemplo, a magnitude do impacto ambiental encontrase grandemente influenciado pelo tipo de uso do solo que é convertido em cultivos energéticos. A analise integrada do balanço de GEE da produção de cultivos bioenergéticos e da intensificação do setor agrícola mostra que pode ser alcançado um elevado potencial de redução das emissões de GEE quando o uso agrícola do solo é intensificado e a produção de cultivos bioenergéticos é implementada. Elevadas reduções são alcançadas somente se todo o setor agrícola for gerido de forma mais sustentável. Comparação global de inúmeros resultados quantitativos Na Tabela 7.2 é apresentada uma visão geral dos impactos da produção de bioenergia sob o ponto de vista de potencialidades, de custo e ambientais para as diferentes configurações geográficas. Pode-se concluir que, nos cenários progressivos quantidades consideráveis de terra podem tornar-se disponíveis para a produção de culturas energéticas em Moçambique e na Ucrânia, sem entrar em conflito com outros usos da terra. Isso poderia resultar em elevadas poupanças de emissões de GEE. No entanto, a baixa e decrescente disponibilidade de terras potenciais no cenário BAU indica uma maior competição por terra no futuro. Deve-se, portanto ressaltar que uma grande escala de produção do setor de bioenergia sustentável só pode ser estabelecida se ele for desenvolvido simultaneamente com um setor mais produtivo e agricolamente sustentável. Isto implica em uma suspensão das tendências atuais. Para Moçambique isto significa uma mudança de uma agricultura de subsistência para a agricultura comercial e da pastoral para os sistemas pecuários mistos. Isso exige mudanças na gestão agrícola (especialmente a implantação de fertilizantes e sementes melhoradas), desenvolvimento de mercados regionais ou internacionais, melhoramento da logística, treinamento e melhores capacidades globais e de governança do setor agrícola. Para a Ucrânia, isso implicaria a reforma terra que facilitaria a agricultura comercial, o alinhamento com padrões internacionais de produtos agrícolas, um sistema econômico que permite a renda nãoagrícola e acesso ao capital e insumos agrícolas

283

SHADES OF GREEN

Tabela 7.2: Visão global de petencialidades, custos e impactos ambientais da produção de Bioenergia para diferentes regiões geográficas. Unidade Holanda Moçambique Ukrânia a Disponibilidade de terras Mha 0.04 - 0.05 7.7 -16.4 0.03 - 31.2 Potencial de produção de bioetanol 1ª geração b de etanol PJ 6 -7 350 - 850 1 - 2370 Potencial de produção de bioenergia do c colheitas lenhosas / herbáceas PJ 7-9 1340 - 3200 3 - 2230 d Custos do bioetanol de 1ª geração €/ GJ 27 - << 14 - << ~11 - << Custos do bioetanol de 1ª geração / pelotas d torrefeita €/ GJ 24 - << 5 - << ~9 - << Impacto ambiental da bioenergia 1ª geração colheitas e -/+ Impacto ambiental da bioenergia colheitas lenhosas / herbáceasc +/+ a Para os Países Baixos, esta inclui apenas a terra atualmente em uso como terras agrícolas e baseia-se nos resultados de de Wit e Faaij (2010). Em Moçambique e na Ucrânia terra disponível também inclui outras terras que compreende principalmente pastagens e matagal (floresta, áreas protegidas, íngreme encostas, etc, são excluídos). b Na Holanda este é o etanol da beterraba de açúcar, na Moçambique este é o ethanol da cana de açúcar, na Ukrânia este e’o ethanol do trigo. c Na Holanda este é o etanol da Miscanthus, na Moçambique este é os pelotas torrefeita de eucalypto, e na Ukrânia este é o etanol de switchgrass. d O custo da produção de etanol na Ucrânia são o custo de produção de etanol, considerando o menor custo de matéria-prima derivado de Wit e Faaij (2010). O custo previsto para Moçambique incluem o transporte a partir da planta de conversão para o porto, a armazenagem e transporte de longa distância, Estes custos são excluídos dos cálculos de custo da produção de etanol na Holanda e Ucrânia. e Os impactos ambientais incluídas na Holanda incluiu emissão de GEE, e de impacto e de água do solo e da biodiversidade. Na Ucrânia, apenas emissão de GEE foram avaliadas. Para Moçambique, os impactos ambientais da produção de bioenergia não foram avaliados nesta tese.

IV.

Qual confiabilidade pode ser obtida utilizando os dados disponíveis e os métodos desenvolvidos neste estudo?

Neste trabalho, uma análise espaço-temporal de desenvolvimento de potencialidades na disponibilidade de terra, e da produção potencial, da perfomance econômica e os impactos ambientais da produção de bioenergia foram considerados. A análise espacial explícita e ex-ante vem com inúmeras incertezas. Mudança de direção do uso da terra Como, em que sentido e em que ritmo os principais impulsionadores da mudança do uso da terra irão se desenvolver em diferentes contextos é incerto. Além disso, a direção, extensão, tipo de correlação e a importância relativa dos condutores para a localização de mudança no uso da terra (como a aptidão agro-ecológica, a acessibilidade, a elasticidade 284

Sumário e Conclusões

de conversão, e de vizinhança) são geralmente difíceis de detectar e podem mudar ao longo tempo, como bem podem eles mesmos mudarem a direção.. Portanto, a validação e calibração do modelo detalhado da mudança do uso espacial da terra é crucial. A avaliação longitudinal através do monitoramento tanto do direcionamento do uso da terra da e suas alterações em diferentes países, bem como em diferentes estágios de desenvolvimento, com diferentes características biofísicas e sócio-econômico poderia vir a melhorar a compreensão das correlações entre direcionamentos e mudanças no uso da terra e como estes irão se desenvolver. Evolução dos custos e viabilidade econômica da produção de bioenergia No longo prazo, os combustíveis fósseis devem ser mais caros o que poderá contribuir para a viabilidade da produção de bioenergia. No entanto, nos últimos anos, os custos de produção de biomassa como matéria-prima foram afetados pelo aumento dos custos dos equipamentos, do diesel, dos fertilizantes e dos defensivos agrícolas. Estes custos podem continuar a aumentar se alinhado com tendências globais históricas. Além disso, os custos de produção em países como Moçambique e Ucrânia são baixos tendo em vista o baixo custo de, por exemplo trabalho e terra. Estes também devem aumentar devido a um aumento da pressão sobre terras e devido ao desenvolvimento econômico. Por outro lado, é esperado que estes aumentos de custos sejam neutralizados por meio de uma gestão mais eficiente (reprodução melhorada por exemplo, as eficiências de conversão superiores, etc.). Competitividade no uso da terra A produção de culturas de bioenergia não foi incluída como uma classe dinâmica do uso da terra, sendo uma ordem fixa de alocação das classes de uso da terra aplicado à competição por terra ou mudanças de uso indireto da terra não foi modelado. No entanto, na prática, é provável que a produção de culturas de bioenergia competirá com funções de de utilização terra para outras áreas mais adequadas. Se a competição entre culturas energéticas e outros usos da terra é para ser modelada, a implementação de culturas de bioenergia deve dizer respeito a uma demanda projetada (por exemplo, metas nacionais de mistura de biocombustíveis). Para que possamos modelar a concorrência entre os usos da terra, extensa informação é necessária na evolução do mercado, as elasticidades-preço ,bem como as políticas. No entanto, o ponto de partida fundamental em todas as análises de disponibilidade de terras era que a concorrência pela terra deveria ser evitada, e, portanto, a terra foi excluído se (potencialmente) em uso para funções. A qualidade e a disponibilidade dos dados A explicita avaliação espacial de mudança do uso da terra, dos impactos ambientais e do desempenho econômico requer grandes quantidades de dados (digitais) tais como mapas. 285

SHADES OF GREEN

A disponibilidade e qualidade de dados tem suas limitações e há muitas inconsistências entre várias fontes de dados estatísticos, entre conjuntos de dados espaciais e entre os dados estatísticos puros. Além disso, várias fontes de dados espaciais têm problemas graves de qualidade relacionadas com a resolução, a classificação, consistência e documentação. Além disso, para as análises, os dados de várias fontes são combinados, não sendo assim necessariamente consistentes com os outros. Exemplos disso são os dados sobre a agro-ecológica adequação com os dados do SOC, razão C:N e dados climáticos. Especialmente melhores dados espaciais sobre os principais insumos, tais como de uso corrente da terra, características do solo (por exemplo, sedimentos, SOC, razão C: N, lençol freático), o uso e as condições de pastagens, a adequação agroecológico para várias classes de usos da terra serião fundamentais para melhorar a precisão dos resultados. Além disso, uma maior consistência nos dados estatísticos sobre o uso do solo, níveis de rendimento, produção agrícola, números de animais, etc contribuiria para uma maior fidelidade dos resultados. Uniformidade - Heretogeneidade Na análise das potencialidades, custos e impactos ambientais são assumidos a uniformidade nacional em classes de uso da terra, (evolução) de gestão agrícola e motores de mudança do uso da terra. No entanto, ocorre na prática heterogeneidade espacial considerável nestes parâmetros. Por exemplo, assume-se que a classe uso da terra 'cropland' consistem em uma soma ponderada de todas as culturas produzidas, enquanto que, na prática, a composição de rotações de culturas será diferente para diferentes áreas e localizações. Supõe-se que as práticas agrícolas e a adoção de melhores práticas e os aumentos relacionados em eficiência são uniformes para todos os diferentes tipos de produtores e uniforme para todas as classes agro-ecológicas de adequação. Apenas rendimentos relacionados com as práticas de gestão estão relacionadas, tais como aplicação de fertilizantes ea colheita são variados em relação à adequação agro-ecológica. No entanto, existem atualmente grandes diferenças nas práticas agrícolas e na distribuição espacial de pequena e grande escala de produção. No entanto, por causa da modelagem de mudanças do uso da terra em escala nacional, a média de práticas agrícolas e os aumentos médios em eficiência é assumido para todo o país. Relação de impactos Os cálculos dos impactos ambientais baseiam-se numa vasta gama de parâmetros de insumos. Como todos os impactos estão relacionados com a funcionalidade do ecossistema, eles são fortemente interligados uns com os outros. Em alguns casos, os impactos reforçam-se mutuamente e em outros casos, compensações entre impactos pode ocorrer. No entanto, é complicado para provar causalidade e para quantificar as relações. Os impactos ambientais podem ser modelados de forma mais precisa quando a 286

Sumário e Conclusões

modelagem é mais relacionada à processo embasado ao inves de fazer uso de valores predefinidos e quando os melhores dados estão disponíveis. Além disso, avaliar a significância do impacto causado é complicado, tanto na relação dose-resposta , tal como os atuais danos causados pelo tempo, localização e dimensão. Análise de incertezas A abordagem demonstrada nesta tese fornece possibilidades para lidar com incertezas dentro da modelagem espaço-temporal dos potenciais custos e impactos ambientais. As análises de sensibilidade mostram incertezas em que ponto os resultados de entrada exercem alterações nos dados finais. O cenário abordado neste trabalho permite que a modelagem das principais premissas divergentes, sendo estas os motores da mudança no uso da terra. O modelo PLUC pode lidar com a incerteza, permitindo a análise de Monte Carlo em parâmetros estocásticos de entrada. O desenvolvimento na mudança potencial do uso da terra, os custos e impactos ambientais é calculada com um nível de tamanho 1 km2 de grade, mas os fatores acima mencionados limitam a precisão dos resultados. No entanto, os dados atualmente existentes são considerados para ter precisão suficiente para distinguir padrões e pontos de mudança do uso da terra, viabilidade econômica e desempenho econômico. Portanto, em geral, os resultados podem ser usados para uma primeira triagem para identificar ‘go’ e áreas ‘no-go'. No entanto, se a abordagem apresentada tem como pressuposto ser usada para controle e certificação da produção de biomassa, melhores dados e uma resolução mais criteriosa para modelação são desejados.

7.5

Recomendações para pesquisas futuras •





A fim de avaliar a dinâmica da competição do uso da terra, a alocação de uso da terra também deve ser com base no desempenho econômico relativo de usos alternativos da terra. Para dar conta da dinâmica de preços, oferta e demanda por biocombustíveis, combustíveis fósseis e de commodities agrícolas, a modelagem de mudanças de uso terra deve ser interligada com o geral e / ou modelos de equilíbrio parcial. A avaliação longitudinal através do monitoramento tanto da da alteração nas diretrizes do uso e padrões de uso da terra poderiam melhorar a compreensão sobre a relação entre direcionadores e mudanças no uso da terra, as quais permitem fazer projeções mais confiáveis de possíveis padrões de seu uso. Mais conhecimento é exigido nas inter-relações entre diferentes impactos ambientais. Além disso, a diferenciação das relações dose-resposta e os limiares para dano relacionadas com o contexto da biofísica, tempo e escala requerem mais investigação. Impactos sobre a água deve vir a ser avaliada em um nível da bacia de água para capturar todos os mecanismos relevantes que determinam se

287

SHADES OF GREEN









7.6

a água é utilizada de forma insustentável ou não. Finalmente, mais de validação e coerência é necessária entre os métodos e indicadores para quantificar e monitorar os impactos sobre a biodiversidade da mudança do uso da terra em diferentes níveis espaciais. Os Impactos socioeconômicos da produção de bioenergia estão fortemente relacionadas ao uso da terra, o desempenho econômico e os impactos ambientais. Esses impactos devem ser avaliados de forma integrada, a fim de identificar as áreas mais adequadas e os meios de produção, e ser capaz de lidar com indicadores mais complexos, tais como segurança alimentar, de forma quantitativa. Uso da terra, o desempenho econômico e os impactos ambientais e sócioeconômicos da bioenergia está diretamente relacionado com a dinâmica de todo o setor agrícola (incluindo a utilização de pastagens ea pecuária). Portanto, os efeitos da introdução de produção de culturas para a bioenergia e os desenvolvimentos das funções de uso da terra devem ser analisados de forma integrada. Nesta tese, o primeiro passo é feito através da inclusão de todo o setor agrícola (incluindo cultivos bioenergéticos) na avaliação espaço-temporal das emissões de GEE. Esta abordagem pode também ser aplicada aos outros impactos. Para maior confiabilidade de uma avaliação ex ante 'ir' e 'não passa »áreas de produção de bioenergia, de melhor qualidade (espacial) de dados em termos de precisão, resolução e documentação necessária. Especialmente dados sobre o uso atual da terra (incluindo pastagens e da intensidade de seu uso), agroecológica e adequação às características do solo e clima são obrigatórios. Biorefinaria de biomassa para várias aplicações, tais como alimentos, rações, fibras, energia e produtos químicos oferece oportunidades para o uso eficiente dos recursos de biomassa. Oportunidades de caminhos inovadores precisam ser explorados para melhorar a eficiência global e desempenho econômico das cadeias de fornecimento de biomassa.

Recomendações políticas e para mercados •

Para alcançar os elevados potenciais de implantação de biomassa para energia e ao mesmo tempo evitar iLUC, o aumento da produção de bioenergia, alimentos e rações para animais necessita ser equilibrado com um melhoramento na gestão agrícola. Além disso, é necessária a implementação de quadros políticos e de sustentabilidade para garantir uma boa governação do uso da terra e melhorias na gestão de áreas florestais, agrícolas e produção de animais.

288

Sumário e Conclusões









Existem grandes potenciais de bioenergia em áreas menos desenvolvidas. É necessária uma abordagem mais integrada de desenvolvimento sustentável, a fim de desenvolver o elevado potencial de bioenergia dessas regiões. Para tal, são necessários investimentos não só no setor agrícola, mas também em infraestrutura, educação e desenvolvimento dos mercados. Como os países em desenvolvimento são frequentemente caracterizados por um quadro político e institucional frágeis, a produção de bioenergia nesses países traz responsabilidades acrescidas a outros intervenientes e partes interessadas (produtores, consumidores, organismos de certificação e governos) para que se garanta um desenvolvimento sustentável. Uma análise ex-ante da disponibilidade de terras, da viabilidade econômica e dos impactos ambientais contribui para a identificação das áreas que devem ou não ser permitidas para a produção de bioenergia. Tal permite um planeamento sólido e prudente do uso da terra, do investimento sustentável na capacidade de produção de bioenergia, e da infra-estrutura ao longo do tempo. Pode também ajudar os investidores e decisores políticos a estimarem realisticamente a viabilidade econômica dum projecto, permitindo definir as condições prévias para cumprir com os critérios de sustentabilidade. Tal poderia ajudar a reduzir os riscos de investimento e evitar o fracasso de projectos de larga escala. A certificação avançada de biomassa sustentável a partir de parâmetros verificáveis e quantificáveis requer ferramentas aperfeiçoadas para avaliar o desempenho ambiental e económico e os impactos do uso do solo. A abordagem aqui apresentada pode contribuir para uma análise ex-ante para a certificação da sustentabilidade da produção de bioenergia. No entanto, é ainda necessária uma melhor qualidade de dados e uma resolução de modelação mais fina. Idealmente, as políticas de agricultura, ambiente, energia renovável e desenvolvimento rural necessitam ser alinhadas a fim de desenvolver um setor de bioenergia sustentável, o qual consiste numa componente da governação sustentável da terra em geral.

289

Резюме та висновки

Резюме та висновки

291

SHADES OF GREEN

7.1

Контекст дослідження

Будь-яке суспільство споживає енергетичні ресурси для задоволення своїх основних потреб та для полегшення протікання виробничих процесів (IPCC 2011). На сьогоднішній день 85% від загального обсягу первинних енергоресурсів отримується з викопних видів палива (МЕА 2010; IPCC 2011). Проте, такі енергетичні системи не можуть бути сталими у довгостроковій перспективі через вичерпність викопних ресурсів, нерівномірне їх розподілення, а також, зважаючи на значний внесок викопного палива до антропогенних викидів парникових газів. Заміщення споживання викопного палива шляхом розширення використання відновлювальних джерел енергії є необхідним заходом для розвитку енергетичних систем з дотриманням критеріїв сталого розвитку (IPCC 2007). Наразі внесок біомаси до загального енергоспоживання оцінюється у об’ємі 11 ЕДж (2008 рік) та очікується, що вона і надалі буде відігравати важливу роль (МЕА і ОЕСР, 2011; IPCC 2011). Наприклад, основуючись на аналізі сценаріїв, Міжурядова група експертів зі зміни клімату (IPCC) оцінює, що для задоволення цілей з пом'якшення наслідків викидів ПГ (стабілізація концентрації атмосферного СО2-екв на рівні менше 440 ppm до 2100 року), рівень споживання біомаси в 2050 році має становити 120190 ЕДж (IPCC 2011). Проте, збільшення виробництва і використання біомаси стимулюється не тільки її потенціалом для пом’якшення парникового ефекту (при умові дотримання критеріїв сталості), але і наступними чинниками: порівняно легкою реалізацією біоенергетичних проектів в існуючій енергетичній інфраструктурі; універсальністю біомаси як енергоресурсу; диверсифікацією енергопостачання та подальшим збільшенням енергобезпеки; потенційним внеском в розвиток сільських районів та потенціалом для відновлення виснажених земель. Незважаючи на численні переваги, підвищення рівня використання зазначених культур для виробництва енергії може мати значні негативні соціально-економічні та екологічні наслідки: вирубування лісів, втрату вуглецевих поглиначів, біологічного різноманіття та інших функцій екосистем, переміщення населення і загострення конкуренції за землю, воду та інших факторів виробництва, що, в свою чергу, може привести до підвищення цін на продукти харчування (IPCC 2011). Багато з цих негативних наслідків пов'язані зі зміною в користуванні земельними ресурсами (LUC) (Wicke та ін. 2012 рік.). Тому, для досягнення високого рівня розгортання біоенергетики, необхідно уникати конкуренції між продуктами харчування, кормами та паливом - а значить, і непрямої зміни у землекористуванні (iLUC ). Досягнення такого результату можливе шляхом врівноваження зростання виробництва біомаси для енергетичних цілей та покращень в управлінні сільським 292

Резюме та висновки

господарством (Dornburg та ін. 2010; Wicke та ін. 2012 рік.). Крім того, основні екологічні проблеми повинні вирішуватися шляхом вибору відповідних біоенергетичних систем та застосування адекватного планування землекористування (Dornburg і співавт. 2010). Ефективність виробництва та використання біомаси залежить від конкретного регіону (IPCC 2011), а різноманітні впливи відбуваються як на місцевому, регіональному так і на глобальному рівні (van Dam і співавт., 2010). Реалізація ефективних основ сталості, наприклад, шляхом розробки схем сертифікації, може пом'якшити негативний вплив на довкілля та дозволяє, одночасно, внести вклад до реалізації різноманітних цілей сталого розвитку. Кодекси поведінки та протоколи за критеріями сталості були розроблені, та наразі розробляються на декількох рівнях. В даний час ключовою проблемою як для учасників ринку, так і для уряду, є наступне запитання: як такі критерії можуть бути виконані на практиці, та яким чином можна якісно та надійно перевірити кількісні характеристики всіх негативних наслідків. Для забезпечення ефективної сертифікації, раціонального планування стійких інвестицій в майбутньому і належного управління в галузі землекористування та сільського господарства (що напряму пов’язано із зростанням виробництва біомаси та подальшим її використанням для енергетичних цілей) має бути проведено покращення методики «просторово-орієнтованої» оцінки потенціалу та впливів на навколишнє середовище.

7.2

Мета та завдання дослідження

Метою даної роботи є дослідження того, як потенціал, витрати та екологічні наслідки виробництва біопалива можуть бути оцінені з врахуванням уникнення ilUC і просторово-часової мінливості біофізичних та соціально-економічних властивостей. В результаті розглядалися наступні запитання: I.

II.

Як потенційна наявність землі для вирощування енергетичних культур може бути оцінена просторово- та часово-орієнтовано , враховуючи, що слід уникати непрямої зміни у землекористуванні, а, отже, приймаючи до уваги розвиток його інших функцій? Як за допомогою просторово- та часово-орієнтованого методу оцінити економічну життєздатність та екологічні наслідки виробництва біопалива, враховуючи: розташування конкурентів – інших користувачів сільськогосподарських земель, вартість виробництва біомаси і логістику поставок, та вплив на викиди парникових газів, якість ґрунту, води та біорозмаїття?

293

SHADES OF GREEN

III. IV.

Який потенціал, економічні показники та екологічні наслідки виробництва біомаси за різних умов? Яка достовірність порівняння результатів, отриманих за допомогою доступних даних і методів, розроблених в даному дослідженні?

Вищеперераховані запитання розглядаються в главах 2-6 роботи. У главах 2 та 3 шляхом «просторово-орієнтованої» оцінки економічної життєздатності та потенційних екологічних наслідків для ланцюжка виробництва біомаси в регіонах дається відповідь на перше запитання. Дана оцінка виконується з врахуванням просторових змін в агроекологічній придатності, поточного використання землі та інших біофізичних факторів. Методики ГІС, які були розроблені в даних розділах, дозволяють «просторово-орієнтовану», але статичну оцінку витрат та екологічних наслідків. Північ Нідерландів була обрана як навчальний регіон, зважаючи на високу конкуренцію за землю та пов'язану з цим інтенсивність її використання, а також через наявність детальних просторових даних. У розділі 4, враховуючи місцеві рушійні сили змін у землекористуванні, була розроблена нова модель (PLUC), яка в подальшому була використана для оцінки наявності земель для вирощування біоенергетичних культур. Розроблена модель дозволяє виконати просторово- та часово-орієнтовану оцінку наявності земель для вирощування біоенергетичних культур і полегшує просторово-часову оцінку невизначеності на основі стохастичних вхідних параметрів. Згодом, в розділі 5, були проаналізовані зміни у вартості сировини та логістика ланцюгів поставок біомаси. Даний аналіз був проведений з врахуванням потенційної наявності вільних земель та доступних технологічних даних. Поєднання просторовочасової моделі землекористування, просторово-орієнтованих витрат на виробництво біомаси та на логістику, а також тимчасових витрат, дозволяють виконати просторово-часову оцінку загальної вартості енергопостачання за рахунок біомаси. У главі 6 розроблена модель була адаптована для умов України та розширена модулем по викидам парникових газів, з тим, щоб в подальшому проаналізувати можливі зміни у викидах ПГ всього сільськогосподарського сектора, включаючи впровадження вирощування біоенергетичних культур та інтенсифікацію сектора сільського господарства. Зв'язок між динамічним модулем розрахунку викидів парникових газів та моделі змін у землекористуванні дозволяє провести просторовочасову оцінку впливу викидів парникових газів усього числа змін у землекористуванні. Зважаючи на високий потенціал виробництва енергії з біомаси (даний потенціал спричинений низькою щільністю населення, сприятливими умовами для виробництва біомаси, та через різноманітність екологічних та соціально-економічних умов), Мозамбік та Україна були обрані у якості навчального прикладу. У розділах 2, 294

Резюме та висновки

3, 5 та 6 проводиться аналіз вирощування типових енергетичних культур для виробництва біопалив першого та другого покоління. У (майже) всіх дослідницьких прикладах ці енергетичні культури розцінюються як сировина для виробництва біоетанолу. Це дозволяє забезпечити порівняння відмінностей в потенціалах, витратах та впливах на навколишнє середовище типових біопалив першого і другого покоління. Складність і рівень інтеграції методології, що розробляється в даній дисертаційній роботі, збільшується від розділу до розділу, та розвивається від просторовоорієнтованого та статичного моделювання (у розділі 2: економічні показники, у розділі 3: впливи на довкілля) до просторово-часового та динамічного моделювання (в розділі 4: зміни у землекористуванні, у розділі 5: зміна вартостей поставок). У главі 6 просторово-часове моделювання змін у землекористуванні інтегрується з динамічним моделюванням викидів парникових викидів, що дозволяє виконати комплексну просторово-часову оцінку впливів. У всіх розділах наголошується на обмеженості методів і даних, що використовуються у відповідних оцінках. Таблиця 7.1 представляє огляд відповідних глав та дослідницьких запитань, що вирішуються. Таблиця 7.1: Огляд досліджуються

розміщення глав дисертаційної роботи та запитань, що в них

Глава

2 3 4 5 6

Потенціал, просторове розподілення та економічні показники місцевих ланцюгів біомаси Просторова зміна впливів місцевих ланцюжків біомаси на навколишнє середовище Просторово-часове моделювання землекористування для оцінки наявності земель для вирощування енергетичних культур Просторово-часова характеристика вартості виробництва енергії з біомаси Повний просторово-часовий аналіз сільськогосподарського землекористування, потенціалу виробництва біомаси та відповідних балансів парникових газів

7.3

Дослідницькі запитання I II III

IV

































Резюме основних результатів

У розділі 2 шляхом аналізу просторових змін в економічній доцільності виробництва етанолу з міскантуса та цукрових буряків на півночі Нідерландів розглянуті питання II, III та IV. Аналіз конкурентоспроможності біоенергетичних культур проводиться за рахунок: розрахунку показника чистої приведеної вартості (ЧПВ) для багаторічних 295

SHADES OF GREEN

культур; поточної схеми сівозмін; схеми сівозмін, яка включає в себе додаткові роки для вирощування цукрових буряків; порівняння собівартості виробництва біоетанолу з середньою вартістю бензину. Сучасне використання земель та придатність ґрунтів для вирощування поточних та біоенергетичних культур були нанесені з використанням географічних інформаційних систем (ГІС). Для позначення на карті місць, з найбільшою імовірністю змін у землекористуванні, був використаний метод просторового розподілу економічної рентабельності. Собівартість виробництва біоетанолу включає витрати, пов'язані з вирощуванням, збиранням, транспортуванням сировини, та її перетворенням в етанол. ЧПВ та вартість виробництва сировини були розраховані для семи класів придатності ґрунтів. Результати показують високе просторове коливання як у вартості, так і у рентабельності виробництва біомаси в порівнянні зі звичайними сільськогосподарськими культурами. За поточних ринкових цін біоенергетичні культури не можуть конкурувати із традиційними при їх вирощуванні на ґрунтах, які класифікуються як "придатні". На менш придатних ґрунтах рентабельність культур, що вирощуються за інтенсивної технології, низька, і багаторічні культури досягають кращих показників ЧПВ, ніж звичайні сівозміни. Наші результати показали, що мінімальна собівартість виробництва сировини становить 5,4 €/ГДж для міскантуса та 9,7 €/ГДж для цукрових буряків залежно від придатності ґрунтів. Виходячи з показників собівартості, етанол з міскантуса (24 €/ГДж) є кращим варіантом, ніж етанол, вироблений з цукрових буряків (27 €/ГДж), проте, виробництво біоетанолу з власних, місцевих сільськогосподарських культур та за теперішніх умов не може конкурувати з отриманням бензину (12,34 €/ГДж). Проте, при збільшенні цін на нафту, або при появі відповідних технологічних знань та при умові підвищення економічної ефективності комплексного виробництва біоетанолу, теплової та електричної енергії, конкурентоспроможність даного виду пального може бути підвищена. У даному розділі міститься загальна методологія по виявленню перспективних місць для вирощування енергетичних культур з економічної точки зору, з урахуванням просторових змін у поточному землекористуванні та біофізичних факторів, таких як властивості ґрунту та наявність води. Глава 3 містить запитання II, III і IV. В даній главі кількісно оцінюються просторові коливання можливого екологічного впливу при виробництві енергетичних культур. Як приклад розглядається вирощування цукрових буряків та міскантуса для виробництва біоетанолу на півночі Нідерландів. До факторів впливу на довкілля відносяться: викиди парникових газів (ПГ) (протягом життєвого циклу та пов'язані з прямою зміною у землекористуванні), якість ґрунту, кількість і якість води та біорозмаїття. Для кожного фактору на підставі огляду великої кількості літератури були обрані відповідні показники та методи. Дані показники при необхідності їх 296

Резюме та висновки

просторово-орієнтованого використання були адаптовані. Просторові коливання факторів впливу на навколишнє середовище, що пов'язані з просторовою неоднорідністю їх фізичних властивостей, оцінюються за допомогою географічної інформаційної системи (ГІС). Розглянуті приклади показують, що існують великі просторові відмінності екологічних наслідків впровадження біоенергетичних культур. Загалом, при вирощуванні цукрових буряків виникає низка негативних факторів, що впливають на довкілля, особливо на пасовищах. У цих областях викиди парникових -1 газів, що пов’язані зі зміною у землекористуванні становлять 148 кг∙ГДжетанол , а -1 ризик виникнення ерозії ґрунтів збільшується до 9 т∙га . Крім того, в цих районах існує високий ризик втрати біорозмаїття. Серед позитивних факторів впливу на -1 навколишнє середовище в цих областях: зниження на 75 мг∙л концентрації NO3 в ґрунтових водах та зниження сезонного дефіциту води на 100 мм. При переведенні ріллі на вирощування міскантуса викиди парникових газів можуть бути скорочені на -1 -1 159 кг∙ГДжетанол , а ризик виникнення ерозії ґрунту може бути скорочено на 4 т∙га , -1 концентрація NO3 знижується на 53 мг∙л , робиться позитивний вплив на біологічне різноманіття, проте, з іншого боку, на 150 мм збільшується сезонне виснаження води. Західні вологі пасовища є областями з найбільшими негативними наслідками для обох культур. Об’єднані результати просторових оцінок показують, що існує декілька компромісів між факторами впливу на навколишнє середовище: немає таких ділянок, на яких не було б виявлено жодного негативного впливу. Виконана оцінка демонструє межі для визначення потенційних негативних екологічних наслідків виробництва енергії з сільськогосподарських культур та райони, де вирощування енергетичних культур мають невеликий або позитивний вплив на навколишнє середовище. Глава 4 звертається до I, III і IV запитань. У главі розробляється модель для просторово-орієнтованої оцінки майбутніх змін в наявності земель для вирощування біоенергетичних культур. Увага акцентується на недопущенні конкуренції з іншими функціями землекористування, і, отже, на уникненні непрямої зміни землекористування. У роботі визначені та включені до моделювання: місцеві механізми, що стимулюють зміни у землекористуванні для розвитку інших його функцій, таких як: вирощування продуктів харчування, тваринництво та виробництва матеріальних ресурсів; та невизначеності в цих механізмах. Ця просторово-часова растрова модель змін у землекористуванні (PLUC) демонструється на прикладі зростання доступності земель для вирощування біоенергетичних культур в Мозамбіку на протязі 2005-2030 рр.. Основуючись на прогнозах по зміні чисельності населення, дієти, валового внутрішнього продукту та рівня самозабезпечення, проводилась оцінка розвитку основних механізмів використання сільськогосподарських земель, попиту на продукти харчування, тваринні продукти та 297

SHADES OF GREEN

матеріали. Було розроблено два сценарії розвитку: звичайний та прогресивний. Розміщення земельних ділянок основується на конкретному наборі класових факторів, що визначають їх придатність для тих чи інших цілей. Динаміка змін у 2 землекористуванні була нанесена на мапу з сіткою 1 км для кожного року до 2030 року. 7,7 млн. га земель за звичайним, та 16,4 млн. га за прогресивним сценарієм можуть стати доступними для виробництва енергетичних культур в 2030 році. На основі аналізу Монте-Карло, в наявності є 95% довірчого інтервалу від визначеної кількості земель та було знайдено просторово-орієнтовану ймовірність доступних земель. Принцип висхідного аналізу, кількість динамічних землекористувачів, різноманітний перелік рушійних сил змін у землекористуванні та факторів придатності земель, а також можливість моделювання невизначеності дозволяє значно покращити інтегральне моделювання наявності землі для біоенергетичних цілей. Крім того, вона надає можливість вивчення передумов для досягнення високого рівня розвитку біоенергетики, уникаючи при цьому iLUC. У главі 5 шляхом просторово-орієнтованої оцінки того, як вартість енергії із біомаси та її потенціал змінюється з часом, розглядаються запитання II, III і IV. Оцінка вартості та потенціалу біомаси виконується на основі показників доступності земель, придатності земель, які є та можуть стати доступними, з врахуванням вартості виробництва енергетичних культур, відстані транспортування сировини на переробний завод, вартості перетворення, відстані транспортування від заводу до порту та вартості міжнародного судноплавства. У якості дослідницького прикладу розглядаються ланцюжки поставки обпалених гранул з евкаліпту та біоетанолу з цукрових буряків у Мозамбіку. Зміни у доступності земельних ресурсів для вирощування енергетичних культур у Мозамбіку отримуються на основі висновків до глави 5. Результати показують широкі просторові зміни у вартості поставок біомаси, що є підсумком просторових змін у вартостях виробництва сировини, первинних та вторинних транспортних витрат. Найбільш перспективні зони для вирощування евкаліпта та цукрової тростини зосереджені в центрально-південних, центральних та північно-східних частинах Мозамбіку. Дані регіони характеризуються відносно сприятливими агроекологічними умовами, можливістю отримання достатньої кількістю сировини (для задоволення початкових умов технології перетворення), та наявністю відповідної інфраструктури. Для прогресивного сценарію, загальний розрахунковий потенціал по виробництву обпалених гранул з евкаліпту становить 3200 ПДж в 2030 році, з яких 2500 ПДж можуть експортуватися в Європу по ціні, нижче рівня ринкової – від 8 €/ГДж; для етанолу з цукрової тростини загальний потенціал оцінюється в 850 ПДж, з яких 500 ПДж можуть бути експортовані за ціною, нижче ринкової – від 30 €/ГДж. Ключовим фактором економічної ефективності виробництва є його вдале розміщення. Даний фактор є особливо актуальним для 298

Резюме та висновки

країн з високою неоднорідністю агроекологічної придатності та з погано розвиненою інфраструктурою. Підвищення кількості районів для економічно привабливого виробництва енергії із біомаси можливе за рахунок зниження собівартості транспортування, що в свою чергу спричиняється покращенням автомобільної та залізничної інфраструктури. Дане дослідження демонструє підхід, який дозволяє проводити оцінку потенціалу та вартості енергії із біомаси у часі та просторовоорієнтованим чином. Оскільки екологічні та соціально-економічні наслідки біоенергетичних ланцюгів широко пов'язані з біофізичними та соціальноекономічними умовами розміщення виробництва, просторово-орієнтована оцінка потенціалу виробництва енергії із біомаси є вдалим підходом для проектування, оптимізації та аналізу впливів та сталості ланцюжків біоенергетики. У главі 6 на основі просторово-орієнтованого аналізу розвитку біоенергетичного потенціалу та потенціалу скорочення викидів ПГ в Україні на протязі 2010-2030 років (враховуючи розвиток та викиди від інших, сільськогосподарських функцій землекористування) розглядаються запитання I, II, III і IV. Зміна у потребі земельних ресурсів для виробництва продуктів харчування та кормів аналізується просторово, на щорічній основі та з використанням моделі PLUC. Шляхом адаптації динамічних класів землекористування, факторів придатності та їх характеристик до української ситуації, модель була спеціально пристосована для умов України. Для періоду 20102030 рр. проводилась оцінка двох сценаріїв: звичайний, в якому зберігаються нинішні тенденції продуктивності праці; та прогресивний сценарій, який прогнозує збіжність рівня врожайності культур в Україні та в країнах Західної Європи. За прогресивним сценарієм 32,2 млн. га землі можуть стати доступними для виробництва енергетичних культур до 2030 року. Прогнозований рівень розвитку землекористування слугує в якості вхідних даних для просторово-часового балансу парникових газів (CO2, N2O і CH4). Викиди пов'язані зі змінами у галузі управління і використання земель, а зниження викидів парникових газів, в свою чергу, пов’язані із заміщенням викопних видів палива біоетанолом, виготовленим із пшениці та проса. Просторово-часовий модуль балансу ПГ генерує просторово-орієнтовані карти 2 (розширення 1 км ) окремих викидів парникових газів для кожного року. При вирощування проса або, коли на занедбаних сільськогосподарських землях відновлюється природна рослинність, досягається високий рівень зв'язування вуглецю. Ділянки землі, що звільняються в основному розташовані в північній, східній та південній частинах України. При вирощування на покинутих сільськогосподарських землях проса або пшениці, сумарні викиди N2O збільшуються. Особливо це явище спостерігається в районах, де вирощування зазначених культур поширюється на землі пасовищ. Результати показують, що при розвитку за прогресивним сценарієм, сукупний баланс ПГ у 2030 році може становити: -0,8 млрд. т СО2-екв – при 299

SHADES OF GREEN

вирощування пшениці, та -3,9 млрд. т СО2-екв – при вирощуванні проса. Негативний баланс при вирощування пшениці обумовлений, також, відновленням природної рослинності на занедбаних сільськогосподарських угіддях, що виключає виробництва на цих землях енергетичних культур. Коли приймаються додаткові заходи щодо пом’якшення викидів ПГ (наприклад: зниження обробки ґрунту, внесення більшої кількості органіки та використання досконаліших добрив), з метою скорочення викидів CO2 і N2O сільськогосподарського походження, сукупний баланс парникових газів у 2030 році може навіть покращитися: до -2,6 млрд. т СО2-екв – при вирощування пшениці, та -5,1 млрд. т СО2-екв – при вирощуванні проса. У випадку використання звільненої землі для повторного відновлення природної рослинності, значну кількість вуглецю буде акумульовано у вигляді біомаси та ґрунтового органічного вуглецю. Це може спричинити досягнення балансу ПГ на рівні -4,4 млрд. т СО2-екв в 2030 році. Однак, це зв'язування вуглецю може бути проведено лише один раз, на відміну від зниження за рахунок вирощування енергетичних культур. В рамках прогресивного сценарію та терміном до 2100 року, сукупні наслідки зменшення викидів ПГ оцінюються в: ± -5,5 млрд. т СО2-екв при відновленні природної рослинності; -6 млрд. т СО2-екв – при вирощуванні пшениці; та -15 млрд. т СО2-екв – при вирощуванні проса. Просторово-часовий модуль балансу ПГ в поєднанні з моделлю PLUC дозволяє просторово-орієнтоване та динамічне моделювання викидів ПГ, спричинених змінами в землекористуванні та управлінні при вирощуванні енергетичних культур. На основі розробленої методології та висновків до розділів 2 – 6 були надані відповіді на головні питання дослідження, представлені рекомендації та визначені напрямки подальших досліджень.

7.4

Основні висновки I.

Як потенційна наявність землі для вирощування енергетичних культур може бути оцінена просторово- та часово-орієнтовано, враховуючи розвиток інших функцій землекористування та уникаючи непрямої зміни призначення земель?

У даному дослідженні була розроблена нова модель для врахування змін у землекористуванні (PLUC), яка призначається для оцінки розвитку наявності земель 2 для вирощування енергетичних культур на детальному просторовому рівні (1 км ), з урахуванням динаміки і невизначеності ключових факторів для зміни землекористування.

300

Резюме та висновки

Основними факторами підвищення попиту на сільськогосподарські продукти в регіоні є чисельність населенні, ВВП, раціон, використання матеріалів на душу населення та рівень самозабезпечення. Кількість землі, що є необхідною для задоволення загальних потреб, залежить від ефективності сільського господарства і агроекологічної придатності виробничої області. Через невизначеність у виникненні механізмів змін у землекористуванні, для вивчення рушійних сил у довгостроковій перспективі може бути використаний сценарний підхід. Загальна кількість землі, що необхідна для задоволення попиту на продукти харчування, деревину та продукти тваринного походження, безпосередньо пов'язана із продуктивністю розташування певного класу землекористування (наприклад орні землі, пасовища, ліси). Виділення землі до класу динамічного землекористування базується на придатності даного місця до даного класу. Це визначається поєднанням чинників просторовоорієнованої придатності (пов'язані з агроекологічними властивостями), досяжності, еластичності перетворення та суміжності з іншими ділянками. Для кожного фактора придатності мають бути визначені напрямок, вид і ступінь взаємозв’язку. Фактори придатності, їх характеристики та вагомість є специфічними для кожного регіону та класу землекористування. Виключенню підлягають ділянки, що не підходять (наприклад круті схили) або не дозволяються для перетворення в сільськогосподарські угіддя (наприклад природоохоронні території). Відправною точкою для розподілу землекористування слугує карта його поточного стану, калібрована статистикою по потребі у їжі, кормах та матеріалах, продуктивністю сільського господарства, а також розподілом класів землекористування по агроекологічній придатності. Земля виділяється під конкретний клас землекористування в часі інтервалом в один рік. Загальне розподілення землекористування для кожного часового кроку закінчується тоді, коли всі класи будуть розміщені, а загальне виробництво на цих землях досягне значення повної потреби у всіх ресурсах для конкретного року. Моделювання включає в себе контур зворотного зв'язку: карти землекористування, що були отримані в результат розподілення для часу t, слугують в якості початкових даних для розміщення в момент часу t+1. Результатом даного типу динамічного моделювання змін у 2 землекористуванні є побудова карт детального просторового рівня (1 км ), на яких позначено наявність земель під вирощування енергетичних культур для кожного року періоду моделювання. Основною перевагою даної моделі є її здатність опрацьовувати випадкові вхідні дані. Це дає можливість використовувати просторово-часовий аналіз Монте-Карло (МК), який оцінює розповсюдження невизначеностей. PLUC випадковим чином моделює часові ряди (наприклад, потреба у сільськогосподарських культурах та продуктивність сільського господарства), просторові параметри вводу (наприклад, 301

SHADES OF GREEN

щільність населення і продуктивності), і характеристики факторів придатності (наприклад, максимальна відстань до дороги). Випадкове моделювання дозволяє провести аналіз чутливості результатів при невизначеності ключових параметрів. У 2 результаті, на карти наносяться: ймовірність наявності певного місця (сітка 1 км ) в певний відрізок часу (рік). Розроблена у даному дослідженні модель є передовим інструментом для оцінки майбутньої динаміки землекористування та наявності земель для біоенергетичних культур. Застосування сценарного підходу на ключові рушійні сили змін у землекористуванні та використання парадигми харчової першості, допускає виконання оцінки потенціалу біомаси, який може бути реалізований без конкуренції з продуктами харчування та кормами, а також вказати на необхідні умови для реалізації цього потенціалу. Принцип висхідного аналізу, кількість динамічних землекористувачів, різноманітний перелік рушійних сил змін у землекористуванні та факторів придатності земель, а також можливість моделювання невизначеності є кроком вперед до моделювання наявності землі для енергетичних культур. Як врожайність біомаси, витрати на виробництво, логістику, так і вплив на навколишнє середовище тісно пов'язані з місцевими, конкретними біофізичними умовами (агроекологічна придатність, наявність інфраструктури, властивості ґрунту, кліматичні умов і т.д.). Просторово-орієнтована оцінка наявності земель для біоенергетичних культур є важливою передумовою для розробки логістики поставок біомаси, оцінки як потенціалу виробництва енергії та охорони навколишнього середовища, так і соціально-економічних наслідків. Модель наразі є пристосованою та продемонстрованою для умов Мозамбіку та України. Тим не менш, вона являє собою гнучку модель, яка може бути використана для інших країн і регіонів з адаптованими вхідними даними, правилами розподілення земель та характеристиками факторів придатності. II.

Як за допомогою просторово- та часово-орієнтованого методу оцінити економічну життєздатність (вартість виробництва та конкурентні переваги) та екологічні наслідки (викиди ПГ, вплив на якість ґрунтів, внутрішні води та біорозмаїття) виробництва енергії із біомаси?

На ранній стадії цього дослідження були розроблені методи просторово-орієнтованої оцінки економічних і екологічних показників, з урахуванням просторової мінливості біофізичних властивостей. На більш пізній стадії дослідження була розроблена просторово-часова модель для оцінки економічної життєздатності та викидів парникових газів, враховуючи як просторові, так і тимчасові варіації.

302

Резюме та висновки

Економічна ефективність виробництва енергії із біомаси залежить від конкурентоспроможності виробництва енергетичних культур, в порівнянні з іншими варіантами землекористування та конкурентоспроможності виробництва енергії із біомаси, в порівнянні з іншими енергетичними системами. Вартість виробництва енергії із біомаси визначають три ключові чинники: вартість виробництва сировини, витрати на логістику і вартість та ефективність технології перетворення. Вартість заготівлі сировини можна оцінити розрахувавши показник чистої приведеної вартості (NPV) всіх елементів витрат (земля, праця, обладнання) і врожайності біомаси за час існування плантації для її вирощування. Просторові зміни у врожайності, відповідних до неї витрат та конкурентоспроможності можуть бути обчислені шляхом об'єднання карти землекористування та агроекологічних карт придатності для конкретних культур. Просторові зміни у вартості логістики біомаси можуть бути проаналізовані з використанням даних про масштаби переробних заводів, щільність біомаси, а також просторової інформації про наявність земель, рівнях врожайності та якості інфраструктури. Собівартість переробки включає в себе інвестиційні, експлуатаційні та витрати на технічне обслуговування, та вартість енергетичних потоків на вході. Витрати на виробництво біомаси та її переробку змінюються в часі у відповідності до наявних технологічних знань. Інтеграція прогнозів у технологічних знаннях до розрахунків затрат, взаємопов'язаних з просторово-часовим моделюванням землекористування, дозволяє виконати просторово та часово-орієнтовану оцінку економічної ефективності ланцюжка поставок енергії із біомаси. Вибір впливів виробництва енергії із біомаси, що аналізуються в даній роботі, на навколишнє середовище проводився основуючись на проблемних областях (стосовно критеріїв сталого розвитку при виробництві енергії із біомаси), які були перераховані кількома національними і міжнародними ініціативами (EC 2009; NEN 2009; RSB 2010). До факторів впливу на довкілля відносяться: викиди парникових газів (протягом життєвого циклу та пов'язаних із змінами у землекористуванні), вплив на якість ґрунтів, води та біорозмаїття. Дані фактори можуть бути кількісно і просторово-орієнтовано визначені шляхом розробки/адаптації існуючих методологій для просторового детального аналізу. Оцінка впливу на водні ресурси проводилась з використанням простого водного балансу та з урахуванням детальних просторових даних по змінам у землекористуванні, ефективних опадів і сукупної випаровуваності. Оцінювання впливу на ґрунти проводилось з використанням Рівняння Вітрової Ерозії та виконанням обліку просторових змін у землекористуванні, характеристиках рослинності, ґрунту, вітру, опадів і температури. Біорозмаїття досліджувалось за допомогою показників середньої відносної чисельності видів, високої природної цінності, та з врахуванням змін у землекористуванні та управлінні, розподілі природних територій та видів, що перебувають під загрозою зникнення. 303

SHADES OF GREEN

На ранній стадії для оцінки викидів парникових газів та впливу на якість води, що пов’язані із змінами у землекористуванні, була використана модель Miterra (Lesschen 2008;. Velthof та ін. 2009). Ця детермінована, статична та векторна модель, яка імітує баланс азоту і фосфору, викидів NH3, N2O, NOx, CH4, вилуговування N, концентрацію NO3 в ґрунтових водах, зміни запасів вуглецю у ґрунті та біомасі, що пов'язані зі змінами в землекористуванні та управлінні. На пізнішому етапі для оцінки 2 викидів парникових газів динамічно та на просторово-детальному рівні (1 км ) була розроблена нова модель. Розрахунок балансу N проводиться, в першу чергу, на основі методів, які були розроблені Lesschen (2008) та Velthof і співавт. (2009), проте адаптовані для растрових розрахунків. На додаток, замість використання фактору викидів N2O за замовчуванням (як запропоновано IPCC), у модель було впроваджено розроблені Lesschen і співавт. джерело азоту, тип ґрунту та питомі викиди N2O від землекористування, вилуговування; стоковий фактор також були дещо адаптовані. Розрахунок викидів вуглекислого газу виконується на основі методу, запропонованого IPCC (2006). Завдяки інтеграції даного модуля викидів ПГ до просторово-часової моделі землекористування, викиди парникових газів можуть бути розраховані просторово- та часово-орієнтовано. Нова модель здатна врахувати просторову зміну у землекористуванні, рівень прибутковості, характеристики ґрунту, клімату, нахилу, а також тимчасові зміни у землекористуванні, управлінні та рівнях врожайності. Крім того, вона дозволяє здійснювати оцінку найбільш придатних районів для інтенсифікації сільського господарства та вирощування енергетичних культур. Більш того, модель може сприяти оцінці умов, за яких баланс парникових газів може бути оптимізованим, в поєднанні з достатнім виробництвом продуктів харчування та кормів. Розроблені методи та моделі представляють підхід для визначення (апріорі) областей, де реалізація виробництва енергії із біомаси є або може стати економічно привабливою та областей з невеликим негативним або навіть позитивним впливом на довкілля. Порівняння просторового розподілу декількох факторів впливу на навколишнє середовище та економічної привабливості дозволяє знайти оптимальне співвідношення між екологічною та економічною ефективністю. Інтеграція цих моделей і методів дозволяє з економічної та екологічної точки зору виявляти підходящі та заборонені зони для виробництва енергії із біомаси, та місця на яких буде відсутня конкуренція на землю. III.

Який потенціал, економічні показники та виробництва енергії із біомаси за різних умов?

304

екологічні

наслідки

Резюме та висновки

При необхідності уникнення непрямих змін у землекористуванні, кількість енергії, що може бути вироблена із біомаси, залежить від розвитку попиту на інші сільськогосподарські продукти, темпів інтенсифікації сільського господарства, та придатності землі, яка стає доступною для вирощування енергетичних культур. У Нідерландах населення зростає досить низькими темпами, тому, передбачається, що щоденний раціон буде стабілізовано на теперішньому рівні. Якщо припустити, що рівень самозабезпеченості залишається стабільним, то очікується, що загальний попит на продукти харчування та корми з плином часу дещо зросте. Нідерланди характеризуються одним з найбільш ефективних і технологічно передових секторів сільського господарства у Європі (de Wit і співавт. 2011). Таким чином, можливості по підвищенню його ефективності є дещо обмеженими (на відміну від України та Мозамбіку). Це призводить до відносно невеликої кількості земель, що є в наявності для вирощування енергетичних культур – від 41 до 51 тис. га в 2030 році (de Wit і Faaij 2010; Fischer та ін. 2010a; Fischer та ін. 2010b). Враховуючи середню врожайність Міскантусу та цукрових буряків, у 2030 році може бути вироблено від 6 до 9 ПДж етанолу. Так як аналіз щодо конкретного розташування доступних земель не проводився, точної оцінки потенціалу вирощування енергетичних культур не може бути виконано. Вартість виробництва етанолу на основі мінімальної вартості виробництва сировини становить 24 €/ГДж для Місканутсу, та 27 €/ГДж для етанолу з цукрових буряків (з урахуванням ефективності технологій, доступних на короткий термін). При швидко зростаючій чисельності населення та покращенні його щоденного раціону у Мозамбіку швидко зросте і попит на сільськогосподарську продукцію. З іншого боку, через дуже низьку продуктивність поточного сільськогосподарського виробництва, існує великий потенціал для його покращення. При оцінюванні звичайного сценарію, де сучасні тенденції розвитку сільського господарства зберігаються, 7,7 млн. га можуть стати доступними; при прогресивному сценарії, який характеризується значним ростом продуктивності сільського господарства, 16,4 млн. га можуть стати доступними у 2030 році для вирощування енергетичних культур. Якщо вільна земля використовується для вирощування евкаліпту, то за звичайним сценарієм у 2030 році може бути отримано 1340 ПДж гранул, за прогресивним – 3200 ПДж. Коли доступні землі використовуються для вирощування цукрової тростини (сировина для виробництва етанолу), 350 ПДжетанол при звичайному та 850 ПДжетанол при прогресивному сценарії може бути вироблено в 2030 році. Найнижча вартість гранул знаходиться на рівні 5 €/ГДж, виробництва етанолу з цукрової тростини – 14 €/ГДж. Зони з найнижчою вартістю – це області, які характеризуються високою продуктивністю, наявністю достатньої кількості сировини для забезпечення 305

SHADES OF GREEN

номінальної продуктивності заводу перетворення, безпосередньою близькістю інфраструктури та портів. При покращенні транспортної інфраструктури вартість логістики може бути зменшено, та як наслідок більше областей може стати економічно привабливими для виробництва енергії із біомаси. Очікується, що в Україні попит на сільськогосподарську продукцію дещо зросте. Хоча чисельність населення зменшується, а щоденний раціон стабільний, в Україні очікується збільшення рівня експорту. Незважаючи на сприятливі агроекологічні умови, продуктивність сільського господарства низька, що, використовуючи прогалини у продуктивності, передбачає великий потенціал для покращення. За звичайним сценарієм, який характеризується низьким, або взагалі відсутнім рівнем підвищення продуктивності, потенціал доступних земель для виробництва енергії із біомаси є обмеженим (0,03 млн. га). Тим не менш, при розвитку за прогресивним сценарієм, у 2030 році 32,2 млн. га можуть стати доступними для вирощування енергетичних культур. Якщо вільна земля використовується для вирощування проса (сировина для виробництва етанолу), то за звичайним сценарієм у 2030 році може бути отримано 3 ПДжетанол, за прогресивним – 2230 ПДжетанол. При вирощуванні пшениці на вільних землях у 2030 році може бути отримано від 1 до 2370 ПДжетанол. Вартість виробництва етанолу на основі мінімальної вартості сировини для умов України становить від 9 до 11 €/ГДж (розрахунки виконувались Wit and Faaij (2010). Найбільшим потенціалом доступних земель для вирощування енергетичних культур володіє Україна, при розвитку за прогресивним сценарієм. Проте, при розвитку за звичайним сценарієм, потенціал вільних земель є досить мізерним. Практична відсутність земель, придатних для виробництва енергетичних культур, зумовлена тим, що найбільшу частку земельних ресурсів займають сільськогосподарські угіддя (79%), залишок – це ліси (15%) або статичні види землекористування (5%, наприклад, зони збереження, будівельна галузь). Дану частку можна збільшити лише за рахунок інтенсифікації сільського господарства (с/г) та, як наслідок, зменшення земель с/г призначення, необхідних для виробництва продуктів харчування та кормів. На сьогоднішній день в Мозамбіку відносно велика доля вільних земель – 9 млн. га. Окрім поточного використання земель для сільського господарства (20%), лісів (60%) та статичних видів землекористування (9%), приблизно 10% землі може бути використана для вирощування енергетичних культур. Проте, при зростаючому попиті на землю (за рахунок збільшення чисельності населення та , як наслідок, потребі в харчових ресурсах та кормах) та відсутності будь-яких покращень в ефективності с/г виробництва, землі для біоенергетичних культур та лісові райони зменшуються.

306

Резюме та висновки

Через суттєві відмінності у біофізичних характеристиках, соціально-економічних умовах, історичних подіях, стадіях розвитку та конкретних політичних умовах, динаміка змін у землекористуванні для Мозамбіку та України визначається різними рушійними силами, або деякі з них впливають більш або менш суттєво. Очікується, що зважаючи на високу залежність від натурального господарства, відсутності інфраструктури та від місцевих ринків, сільськогосподарські землі в Мозамбіку будуть концентруватися в більш густонаселених районах, поблизу великих міст і близько до дорожньої мережі. В Україні навпаки, менше людей безпосередньо залежить від сільського господарства, доступною є щільна транспортна мережа, і тому триває поступовий зсув убік великих, більш комерційних господарств. Тут, агроекологічна придатність є ключовим фактором для розміщення сільськогосподарських угідь, а щільність населення та досяжність угіддя є менш важливою. Зважаючи на вищеперераховані суттєві відмінності параметри для PLUC моделі (стимули, правила розподілу, фактори придатності, їх характеристики та відносні значення) повинні бути адаптовані для кожного конкретного застосування. Оцінка змін вартості поставок показує, що витрати можуть бути значно знижені шляхом підвищення врожайності (та відповідного зниження вартості виробництва сировини), покращення логістики, а також поліпшення попередньої обробки та технологій перетворення. Комплексний аналіз впливів на довкілля, виконаний для північної частини Нідерландів, показує що вони значно різняться для різних місцевостей. Немає жодної місцевості, в якій започаткування вирощування енергетичних культур призводить тільки до позитивних наслідків – всі види впливів взаємопов’язані між собою. Дослідження показує, що ступінь впливу на довкілля суттєво залежить не тільки від розташування місцевості, але й від виду енергетичної культури. Так, вирощування цукрового буряку має відносно багато негативних впливів, особливо при використанні для цього пасовищ. Для таких земель емісія парникових газів внаслідок зміни землекористування становить 148 кг/ГДжетанол, а ризик ерозії ґрунту зростає до 9 т/га. Також в цих місцевостях є великий ризик втрати біорозмаїття. Позитивним впливом на довкілля в даному випадку є зниження вмісту NO3 в ґрунтових водах на 75 мг/л та зменшення дефіциту сезонної вологи на 100 мм. Коли ріллю використовують для вирощування міскантусу, емісія парникових газів може знизитися на 159 кг/ГДжетанол, а ризик ерозії ґрунту – зменшитися на 4 т/га, концентрація NO3 падає на 53 мг/л, вплив на біорозмаїття – позитивний, обсяг сезонної вологи може збільшитися на 150 мм. Просторові схеми впливів на довкілля в основному пов’язані з просторовими схемами поточного землекористування, тобто сила екологічного впливу суттєво залежить від типу землі, яка була «переведена» під 307

SHADES OF GREEN

вирощування енергетичних культур. Комплексна оцінка балансу парникових газів, пов’язаного з вирощуванням енергетичних культур, показує, що можна досягти значного зниження емісії парникових газів при інтенсифікації землекористування та запровадження виробництва енергетичних культур. Значного зменшення обсягів емісії парникових газів можна досягти тільки за умови сталого розвитку сектору сільського господарства в цілому. Порівняння результатів розрахунків для умов різних країн В таблиці 7.2 представлені результати розрахунку потенціалу виробництва і вартості біопалив а також відповідних впливів на довкілля для умов різних країн. Таблиця 7.2: Результати розрахунку потенціалу виробництва і вартості біопалив а також відповідних впливів на довкілля для умов різних країн Розмірність Нідерланди Мозамбік Україна Наявність земель для вирощування енергетичних культур а млн. га 0.04 - 0.05 7.7 -16.4 0.03 - 31.2 Потенціал виробництва біоетанолу 1-го ПДж 6 -7 350 - 850 1 - 2370 покоління b Потенціал виробництва енергії з деревоподібних/трав’яних культур c ПДж 7-9 1340 - 3200 3 - 2230 Вартість біоетанолу 1-го покоління d €/ ГДж 27 - << 14 - << ~11 - << Вартість енергії з біоетанолу 2-го €/ ГДж 24 - << 5 - << ~9 - << покоління / обпалених гранул d Вплив на довкілля енергетичних культур 1-го покоління e -/+ Вплив на довкілля трав’яних культур e +/+ a Для Нідерландів в цю категорію включено тільки землі, які зараз використовуються для сільськогосподарського виробництва (за даними de Wit і Faaij (2010)). В Мозамбіку і Україні землі, доступні для вирощування енергетичних культур, включають окрім сільськогосподарських земель також інші види земель, в основному, пасовиська і землі під чагарниками (ліси, заповідники, круті схили і т.п. виключені). b Для Нідерландів – біоетанол з цукрових буряків, для Мозамбіку – біоетанол з цукрової тростини, для України – біоетанол з пшениці. c Для Нідерландів – біоетанол з міскантусу, для Мозамбіку – біоетанол з обпалених евкаліптових гранул, для України – біоетанол з проса прутєвидного. d Для України собівартість біоетанолу розраховано виходячи з мінімальної вартості сировини (за даними de Wit і Faaij (2010)). Для Мозамбіку вартість виробництва включає витрати на транспортування з заводу до порту, зберігання і перевезення морем на велику відстань, тоді як для Нідерландів і України ці витрати не включаються до розрахунків. e Для Нідерландів впливи на довкілля включають емісію парникових газів, вплив на ґрунтову вологу і на біорозмаїття. Для України вплив на довкілля включає тільки емісію парникових газів. Для Мозамбіку вплив виробництва енергетичних культур на довкілля в дисертації не розглядався.

Можна зробити висновок, що за прогресивним сценарієм великі площі земель можуть бути доступними для вирощування енергетичних культур у Мозамбіку і 308

Резюме та висновки

Україні без конфлікту з іншими видами землекористування. І це потенційно може призвести до значного зниження викидів парникових газів. Однак, за звичайним сценарієм площі під енергетичні культури є невеликими і навіть зменшуються, тому в майбутньому можлива конкуренція між користувачами землі. Треба наголосити, що досягти високорозвиненого сталого сектору біоенергетики можна за умови, що паралельно з ним буде розвиватися сталий сектор сільського господарства з підвищенням його продуктивності. Це передбачає зміну існуючих тенденцій. Для Мозамбіку це означає перехід від сільського господарства, що забезпечує лише прожитковий мінімум, до сільського господарства на комерційній основі, перехід від чабанського тваринництва до тваринництва зі змішаними системами поводження з поголів’ям худоби. Такі переходи потребують змін у практиці ведення сільськогосподарського виробництва (особливо, використання добрив і насіння високої якості), розвитку регіональних та (між-) національних ринків, вдосконалення логістики, проведення навчань і, загалом, – покращення управління сектором сільського господарства. Для України – це проведення земельної реформи, що полегшить розвиток комерційного сільськогосподарського виробництва, гармонізація національних стандартів на сільгосппродукцію з міжнародними, встановлення економічної системи, яка передбачає додаткові доходи для сільгоспвиробників і дає можливості для інвестування в сільське господарство. IV.

Наскільки надійними є результати, отримані з використанням наявних вихідних даних та методик, розроблених в даній роботі?

В цій дисертаційній роботі проведено просторово-часовий аналіз наявності земель для вирощування енергетичних культур, оцінено потенціал виробництва енергетичних культур, економічну ефективність та вплив на довкілля. Просторовий та попередній аналізи характеризуються багатьма факторами невизначеності. Рушійні сили зміни землекористування Наразі неможливо передбачити як саме і з якою швидкістю в різних умовах основні рушійні сили будуть впливати на зміну землекористування. Також важко визначити порівняну важливість кожної сили, яка впливає на потенційне розташування землі, на якій може відбутися зміна виду користування (агроекологічна придатність, доступність, гнучкість перетворення, існуюче оточення). Ці фактори, як і рушійні сили як такі, можуть мінятися з часом. Тому ключовим моментом роботи є обґрунтування та перевірка детальної просторової моделі зміни землекористування. «Повздовжня» оцінка шляхом моніторингу як рушійних сил зміни землекористування, так і систем землекористування різних країн, що знаходяться на різних ступенях розвитку і мають різні біофізичні і соціально-економічні характеристики може допомоги краще зрозуміти взаємодію між рушійними силами та зміною землекористування 309

SHADES OF GREEN

Ріст цін та економічність виробництва біомаси Очікується, що з плином часу викопні палива будуть дорожчати, що буде мати позитивний вплив на економічні показники біоенергетичних проектів. Однак, останніми роками собівартість виробництва біомаси як палива збільшується за рахунок росту вартості обладнання, цін на дизель, добрива та агрохімікати. Ці ціни можуть продовжити свій ріст у відповідності до світових тенденцій. Собівартість виробництва біомаси в таких країнах як Мозамбік і Україна є низькою завдяки низькій вартості трудових ресурсів та землі. Але можна очікувати її ріст з ростом «навантаження» на землю та розвитком економіки. З іншого боку, можна припустити, що завдяки росту ефективності виробництва (наприклад, в тваринництві та інших галузях) ріст собівартості виробництва біомаси буде обмежений. Конкуренція щодо використання земель В роботі не виконувалося як таке моделювання ситуації конкуренції щодо використання земель та непрямої зміни землекористування. Це пояснюється тим, що енергетичні культури не були виділені в клас землекористування, що має динаміку зміни, а до земель іншого призначення був застосований метод фіксованого порядку розподілення. Хоча на практиці, ймовірно, вирощування енергетичних культур буде конкурувати з іншими видами використання найбільш підходящих земель. В разі моделюванні цієї ситуації конкуренції обсяги вирощування енергетичних культур мають відповідати існуючим прогнозам по споживанню (наприклад, національним цілям по використанню сумішевого пального). Для правильного моделювання конкуренції між різними видами землекористування необхідний великий обсяг інформації по ситуації на ринку, гнучкості цін, політиці розвитку. Однак, відправною точкою в усіх оцінках наявності земель було уникнення конкуренції за землекористування. Тому землі, які (потенційно) мають інше призначення, виключалися з таких, на яких можна вирощувати енергетичні культури. Наявність вихідних даних та їх якість «Просторово-орієнтована» оцінка зміни землекористування, впливів на довкілля та ефективності економічної діяльності вимагає великої кількості (просторових) даних та (цифрових) карт. Наявність та якість (просторових) даних обмежена. Існує багато невідповідностей між різними джерелами статистичних даних, між різними наборами просторових даних, між просторовими та статистичними даними. Крім того, декілька джерел просторових даних мають суттєві недоліки, пов’язані з ступенем розрішення, класифікацією, узгодженістю даних та ін. Для виконання даної роботи було скомбіновано вихідні дані з кількох джерел, причому виявилося, що вони не завжди узгоджувалися між собою. Наприклад, дані по агроекономічним характеристикам не відповідали даним по SOC, по співвідношенню C:N та даним по 310

Резюме та висновки

кліматичним умовам. Для підвищення достовірності результатів розрахунків потрібні вихідні дані більш високої якості. Особливо це стосується таких ключових даних як поточне використання земельних ресурсів, характеристики ґрунтів (наявність осадочних порід, SOC, C:N, рівень підземних вод), стан пасовиськ, використання пасовиськ, агроекологічні характеристики земель різного призначення. Також бажано мати більш точну інформацію по врожайності сільськогосподарських культур, поголів’ю худоби і т.п. Одноманітність - різнорідність В даній роботі потенціал, вартість і впливи на довкілля оцінюються виходячи з однакової класифікації видів землекористування, принципів сільськогосподарського виробництва та рушійних сил зміни землекористування. Однак, на практиці існує значна різнорідність цих параметрів. Наприклад, вважається, що сільськогосподарські землі являють собою «зважену» суму всіх культур, що вирощуються, тоді як в реальності сівозміни є різними для різних регіонів. Вважається, що існуюча практика сільськогосподарського виробництва, впровадження більш прогресивної практики виробництва і відповідний ріст продуктивності є однаковим для всіх видів виробників та для земель всіх типів агроекологічної придатності. Для цих земель змінюються тільки такі показники, що впливають на продуктивність, як обсяги застосованих добрив та технології збору. Наразі існують значні відмінності у практиці сільськогосподарського виробництва та розташуванні крупних та мілких виробників. Однак, для цілей моделювання на загальнонаціональному рівні прийняті осереднені показники практики сільськогосподарського виробництва та росту ефективності. Взаємозв’язок впливів Розрахунок впливів на довкілля ґрунтується на великій кількості вхідних параметрів. Оскільки всі впливи відносяться до функціонування екосистеми, вони тісно пов’язані між собою. В одних випадках впливи підсилюють один одного, в інших – можуть збалансовувати один одного. Однак, довести причинно-наслідковий зв'язок та «виміряти» його – важко. Екологічні впливи можна точніше змоделювати, коли моделювання ґрунтується на показниках конкретних процесів, а не на величинах по замовчуванню, і коли наявні більш точні вихідні дані. Крім того, важко оцінити важливість певного впливу, оскільки взаємозв’язок між масштабністю наслідку і силою впливу та «порогове» значення цієї сили, вище якої спричиняється шкода довкіллю, можуть залежати від місця розташування, часу та обсягів. Аналіз факторів невизначеності Підхід, застосований в даній дисертації, дає можливість враховувати фактори невизначеності при моделюванні потенціалу, вартості та впливів на оточуюче 311

SHADES OF GREEN

середовище. Аналіз чутливості показує наскільки невизначеність вихідних даних впливає на результати. Підхід із застосуванням різних сценаріїв дає можливість змоделювати різні рушійні сили зміни землекористування. Модель PLUC враховує фактори невизначеності через застосування аналізу Монте Кало до імовірнісних 2 вхідних параметрів. Розрахунки виконано на сітці з розміром ланки 1 км , але точність отриманих результатів обмежена вказаними вище факторами. Тим не менш, можна вважати, що наявні вихідні дані є достатньо точними для визначення існуючих систем землекористування та місць з можливою зміною землекористування. Загалом, отримані результати можна використовувати для попередньої оцінки придатності регіонів до впровадження біоенергетичних проектів. Однак при використання представленого підходу до моніторингу та сертифікації процесу виробництва біомаси, потрібні більш точні вихідні дані та більш високе розрішення сітки, по який виконується моделювання.

7.5

Рекомендації щодо подальших досліджень •







Для оцінки різних варіантів землекористування в динаміці необхідно провести їх порівняний економічний аналіз. Для отримання даних по динаміці зміни вартості, можливих обсягів виробництва та споживання біопалив, викопних палив та сільськогосподарської продукції треба застосувати модель зміни землекористування разом з повністю або частково рівноважними моделями. «Повздовжня» оцінка шляхом моніторингу як рушійних сил зміни землекористування, так і систем землекористування, може дати краще розуміння взаємодії цих факторів і допомогти в розробці більш надійних сценаріїв майбутнього землекористування. Необхідно більше інформації по взаємозв’язку між різними факторами впливу на оточуюче середовище. Подальшим дослідженням підлягає питання залежності наслідків від сили впливу, визначення граничної сили, перевищення якої може зашкодити екології. Ретельної оцінки потребує питання сталого використання водних ресурсів. Крім того, необхідне глибоке вивчення впливу зміни землекористування на біорозмаїття. Соціально-економічний вплив впровадження біоенергетичних технологій тісно пов'язаний зі землекористуванням, ефективністю економічної діяльності та впливами на оточуюче середовище. Ці впливи потребують комплексної оцінки для визначення найбільш придатних місць та засобів виробництва, для можливості застосування кількісних показників для таких важливих питань як продовольча безпека.

312

Резюме та висновки







7.6

Землекористування, ефективність економічної діяльності, екологічні та соціально-економічні впливи біоенергетики напряму пов’язані з динамікою розвитку всього сектору сільського господарства (включаючи використання земель пасовиськ та розвиток тваринництва). Тому вплив від вирощування енергетичних культур та наслідки різних варіантів землекористування треба оцінювати комплексно. В даній дисертації зроблено лише перший крок в цьому напрямку шляхом включення всього сектору сільського господарства (в тому числі вирощування енергетичних культур) до просторово-часової оцінки емісії парникових газів. Цей підхід можна застосовувати також до оцінки інших видів впливів. Для підвищення надійності результатів попереднього аналізу щодо вибору місцевостей, придатних для виробництва біомаси, необхідні дані більш високої якості/точності. Особливо це стосується даних по існуючому землекористуванню (включаючи використання пасовиськ), по агроекологічних характеристиках ґрунтах та клімату місцевості. Комплексне використання біомаси – для виробництва продовольства, кормів, матеріалів, енергії та хімічних продуктів – дає можливість її ефективного застосування. Для підвищення загальної ефективності та економічності всього ланцюга постачання біомаси необхідний пошук та впровадження інноваційних підходів.

Рекомендації щодо ринків та політики розвитку •



Збільшення обсягів виробництва енергії з біомаси, продуктів харчування і кормів має бути збалансовано вдосконаленням технологій сільськогосподарського виробництва. Це дозволить розширити енергетичне використання біомаси і уникнути непрямої зміни землекористування. Крім того, необхідно впроваджувати політику сталого розвитку для забезпечення раціонального землекористування, лісокористування і вибору правильних напрямків розвитку тваринництва. Знайдено, що великий біоенергетичний потенціал відповідає менш розвиненим місцевостям. Для реалізації потенціалу в цих регіонах необхідний комплексний підхід до питань сталого розвитку. Такий підхід включає інвестиції не тільки в сільське господарство, але також в освіту, інфраструктуру та розвиток ринків. Оскільки для країн, що розвиваються, часто є характерними слабка політична та інституційна база, виробництво енергії з біомаси в цих країнах накладає велику відповідальність на причетні

313

SHADES OF GREEN







сторони (виробників, споживачів, органи сертифікації, уряд) для забезпечення принципів сталого розвитку. Попередній аналіз наявності земель, економічної життєздатності та екологічного впливу проектів є доброю передумовою для ідентифікації регіонів з сприятливими умовами для реалізації біоенергетичних проектів. Для впровадження таких проектів необхідними є довгострокове планування раціонального землекористування, надійні інвестиції в біоенергетичне обладнання та розвиток інфраструктури. На основі результатів попереднього аналізу інвестори та уповноважені особи можуть зробити висновки щодо економічної життєздатності певного проекту та його відповідності критеріям сталого розвитку. Такий підхід дозволяє знизити інвестиційні ризики та уникнути серйозних помилок при плануванні проектів. Для проведення сертифікації процесу виробництва біомаси на предмет його відповідності критеріям сталості необхідна велика кількість надійних даних економічного та екологічного характеру, в тому числі щодо впливу на землекористування. Запропонований підхід може використовуватися для попереднього аналізу сталості процесу виробництва енергії з біомаси. Для більш глибокого аналізу необхідні результати моделювання з більш високим ступенем розрішення сітки. Дуже бажано, щоби політика в секторі сільського господарства була узгодженою з політикою в сфері екології, відновлюваних джерел енергії та розвитку сільських регіонів. Тільки таким чином можна забезпечити сталий розвиток сектору біоенергетики як однієї з складових частин раціонального управління земельними ресурсами.

314

SHADES OF GREEN

316

Samenvatting en conclusies

Samenvatting en conclusies

317

SHADES OF GREEN

7.1

Onderzoek context

De samenleving heeft energie nodig om te kunnen voorzien in haar basisbehoeften (IPCC 2011). Op dit moment wordt 85% van de totale primaire energie geleverd door fossiele brandstoffen (IEA 2010a; IPCC 2011). Het door fossiele brandstoffen gedomineerde energiesysteem is niet duurzaam door de eindigheid van fossiele grondstoffen, de ongelijke geografische verdeling van grondstoffen en de grote bijdrage van het gebruik van fossiele brandstoffen aan de antropogene broeikasgasemissies. Voor een verduurzaming van het energiesysteem is een toenemende inzet van hernieuwbare energiebronnen als alternatief voor fossiele brandstoffen nodig (IPCC 2007a). Het commerciële gebruik van biomassa voor energieproductie bedroeg in 2008 11 EJ, en het wordt verwacht dat biomassa een belangrijkere rol zal gaan spelen in de toekomstige energievoorziening (IEA and OECD 2011; IPCC 2011). Scenarioanalyses van het Intergovernmental Panel on Climate Change (IPCC) projecteren voor 2050 een inzet van 120-190 EJ welke vereist is om de broeikasgasemissie reductiedoelstelling behorende bij een stabilisatieniveau van de atmosferische CO2-eq concentratie onder de 440 ppm in 2100 te kunnen halen (IPCC 2011). De toename in de productie en het gebruik van bioenergie is echter niet alleen toe te schrijven aan het broeikasgasemissie reductiepotentieel (mits duurzaam geproduceerd), maar ook aan de relatief eenvoudige implementatie in de bestaande energie-infrastructuur, de veelzijdigheid van biomassa als grondstof, de diversificatie van de energiemix en de gerelateerde toename van de energie zekerheid, de potentiële bijdrage aan rurale ontwikkeling en het mogelijke herstel van gedegradeerde gronden. Een sterke toename van de teelt van bio-energiegewassen kan echter ook aanzienlijke negatieve socio-economische- en milieu-impacts hebben zoals ontbossing; verlies van koolstof in bodems en (natuurlijke) vegetatie, verlies van biodiversiteit en andere ecosysteemfuncties en -diensten; verdringing van lokale bewoners, en een toenemende concurrentie om land, water en andere productiefactoren, dat dientengevolge kan leiden tot hogere voedselprijzen (IPCC 2011). Veel van deze ongewenste effecten zijn gekoppeld aan veranderingen in landgebruik (Wicke et al. 2012). Om een grote bijdrage van bio-energie aan de energievoorziening op een duurzame manier te kunnen realiseren, dient de competitie tussen voedsel- , veevoeder- en biobrandstoffenproductie - en daarmee mogelijke indirecte verandering in landgebruik door verdringing - worden voorkomen. Dit kan door de toenemende biomassa productie voor energie te compenseren met verbeteringen in de landbouwefficiëntie (Dornburg et al. 2010; Wicke et al. 2012). Of milieueffecten negatief of positief zijn hangt sterk af van de gekozen gewassen in bio-energie systemen en of adequate ruimtelijke ordening wordt toegepast (Dornburg et al. 2010). De impacts en prestaties van de 318

Samenvatting en conclusies

productie en het gebruik bio-energie zijn regio en locatie specifiek (IPCC 2011), en de effecten zijn waarneembaar op lokaal, regionaal en mondiaal niveau (van Dam et al. 2010b). De implementatie van effectieve duurzaamheidkaders, bijvoorbeeld door het ontwikkelen van certificeringsystemen, zou negatieve milieueffecten kunnen beperken en maakt het tegelijkertijd mogelijk bij te dragen aan meerdere doelstellingen met betrekking tot duurzame ontwikkeling. Inmiddels zijn er wereldwijd een groot aantal initiatieven voor de ontwikkeling van duurzaamheidcriteria en richtlijnen voor duurzame implementatie middels certificering. Op dit moment is het voor zowel de overheid als voor marktpartijen de vraag hoe in de praktijk kan worden voldaan aan dergelijke criteria en hoe de diverse effecten kunnen worden gekwantificeerd op een betrouwbare en controleerbare manier. Voor effectieve certificering, degelijke planning van toekomstige duurzame investeringen in biomassaproductiecapaciteit en een goed bestuur van landgebruik en de agrarische sector, zijn betrouwbare en ruimtelijk expliciete potentieel- en impact analyses vereist.

7.2

Doel en onderzoeksvragen

Het hoofddoel van dit proefschrift is te bepalen hoe potentiëlen, kosten en milieu-impacts van bio-energieproductie in samenhang kunnen worden geëvalueerd, rekening houdend met het vermijden van indirecte veranderingen in landgebruik (ILUC) en de ruimtelijk en temporele variabiliteit van de biofysische en sociaaleconomische context. Daartoe zijn de volgende onderzoeksvragen geformuleerd: I. Hoe kan de potentiële beschikbaarheid van land voor bio-energiegewassen ruimtelijk expliciet en in de tijd worden bepaald, gegeven dat indirecte veranderingen in landgebruik moeten worden voorkomen en derhalve rekening houdend met de ontwikkeling in andere landgebruiksfuncties? II. Hoe kunnen de economische haalbaarheid (de locatie specifieke concurrentie met andere agrarisch grondgebruik, de kosten van de biomassa productie en de logistiek van de biomassa keten) en de milieu-impacts (effecten op broeikasgasemissies, bodem, water en biodiversiteit) van bio-energie productie ruimtelijk en tijdsafhankelijk worden bepaald? III. Wat zijn de potentiëlen, de economische prestaties, en de milieu-impacts van bio-energieproductie in verschillende regio’s in de wereld? IV. Met welke betrouwbaarheid kunnen de potentiëlen, kosten en milieu-impacts, met de beschikbare data en de ontwikkelde methoden worden bepaald? De onderzoeksvragen worden behandeld in de hoofdstukken 1 t/m 6. In de hoofdstukken 2 en 3 worden de economische haalbaarheid en de mogelijke milieueffecten van regionale 319

SHADES OF GREEN

bio-energieketens ruimtelijk expliciet onderzocht, rekening houdend met de ruimtelijke variabiliteit in agro-ecologische geschiktheid, het huidige landgebruik en andere biofysische factoren. De op GIS (Geografisch Informatie Systeem) gebaseerde methoden maken een ruimtelijke expliciete maar statische evaluatie van kosten en milieu-impacts mogelijk. Het noorden van Nederland is geselecteerd als casestudie gebied omdat de regio wordt gekenmerkt door intensief landgebruik. Tevens zijn er gedetailleerde ruimtelijk expliciete data beschikbaar voor deze regio. In Hoofdstuk 4 wordt een nieuw landgebruiksveranderingsmodel (PLUC) gepresenteerd dat is ontwikkeld om de landbeschikbaarheid voor bio-energiegewassen te bepalen, rekening houdend met regiospecifieke factoren die bepalend zijn voor veranderingen in landgebruik. Dit model maakt het mogelijk om de beschikbaarheid van land voor bio-energiegewassen ruimtelijk expliciet en temporeel te kwantificeren en onzekerheidanalyse op basis van stochastische invoerparameters uit te voeren. In Hoofdstuk 5 zijn de ontwikkelingen in de kosten van biomassa teelt en de logistiek van de biomassa productieketen geanalyseerd, gegeven de ontwikkelingen in mogelijke beschikbaarheid van land en technologisch leren (verbeteringen in de teelt en conversietechnologieën). De koppeling tussen het spatiotemporele landgebruiksveranderingsmodel, de bereiking van de logistieke kosten, en de analyse van de kostenprijs ontwikkeling, maakt het mogelijk om de kosten van de bio-energieproductieketen locatiespecifiek en in de tijd te kwantificeren. In hoofdstuk 6 is het landgebruiksveranderingsmodel aangepast voor Oekraïne en uitgebreid met een broeikasgasemissie module. Hiermee kunnen broeikasgasemissies van het gehele landgebruik inclusief de implementatie van bio-energiegewassen en de intensivering van de agrarische sector integraal bepaald worden. Mozambique en Oekraïne zijn geselecteerd als casestudie gebieden omdat ze een groot bio-energie productiepotentieel hebben door hun lage populatiedichtheid, gunstig klimaat voor de teelt van bioenergiegewassen en omdat deze landen gezamenlijk een grote diversiteit in de milieu- en socio-economische condities combineren. In hoofdstuk 2, 3, 5 en 6 zijn de prestaties van typische gewassen voor eerste en tweede generatie biobrandstoffen geanalyseerd. In (bijna) alle gevallen is er vanuit gegaan dat de bio-energiegewassen gebruikt worden voor de productie van bioethanol om zo potentiëlen, kosten en milieu-impacts van typische eerste en tweede generatie biobrandstofopties met elkaar te kunnen vergelijken. De complexiteit en de mate van integratie van de in dit proefschrift ontwikkelde methoden neemt toe van hoofdstuk tot hoofdstuk en evolueert van ruimtelijk expliciete en statische modellering (in Hoofdstuk 2 de economische prestatie en in Hoofdstuk 3 de milieu-impacts) naar spatiotemporele en dynamische modellering (in Hoofdstuk 4 de landgebruiksveranderingen en in Hoofdstuk 5 de ontwikkelingen in productiepotentieel en 320

Samenvatting en conclusies

–kosten). In Hoofdstuk 6 is de spatiotemporele landgebruiksmodellering geïntegreerd met dynamische modellering van broeikasgasemissies wat het mogelijk maakt om deze impacts van bio-energieproductie spatiotemporeel te onderzoeken. In alle hoofdstukken worden de beperkingen van de toegepaste methode en de gebruikte data besproken. Tabel 7.1 geeft een overzicht van de hoofdstukken en de onderzoeksvragen die daarin behandeld worden. Tabel 7.1: Overzicht van de thesis hoofdstukken en de onderzoeksvragen die daarin behandeld worden. Hoofdstuk 2 3 4 5 6

7.3

Potentieel, ruimtelijke distributie en economisch resultaat van regionale bio-energie ketens Ruimtelijke variatie in milieu-impacts van regionale biomassa ketens Ruimtelijk expliciete en dynamische landgebruikmodellering voor de evaluatie van land beschikbaarheid voor bio-energiegewassen. Spatiotemporele kosten-aanbodcurves van bio-energie productie Integrale spatiotemporele analyse van agrarisch landgebruik, bioenergie productie potentiëlen en broeikasgasbalansen

Onderzoeksvragen I II III

IV

































Samenvatting van de resultaten

Hoofdstuk 2 adresseert onderzoeksvraag II, III en IV door de ruimtelijke variatie in de economische prestaties van ethanol productie uit Miscanthus en suikerbiet in Noord Nederland te analyseren. Het concurrentievermogen van bio-energiegewassen is onderzocht door de Netto Contante Waarde (NCW) van meerjarige gewassen; huidige gewasrotaties; en gewasrotaties met een extra aandeel suikerbiet met elkaar te vergelijken en door de productiekosten van bioethanol te vergelijken met gemiddelde benzinekosten. Het huidige landgebruik en de bodemgeschiktheid voor conventionele en bio-energiegewassen zijn in kaart gebracht met behulp van een Geografisch Informatie Systeem (GIS). De ruimtelijke distributie in economische rentabiliteit werd gebruikt om aan te geven waar landgebruikveranderingen het meest waarschijnlijk optreden. De productiekosten van ethanol bestaan uit kosten voor de gewasproductie, oogst, transport en de conversie naar ethanol. De NCWs en kosten van biomassaproductie zijn berekend voor 7 bodem klassen. De resultaten laten een grote ruimtelijke variatie zien in zowel de kosten van biomassa als in de rentabiliteit van biomassaproductie vergeleken met conventionele landbouwgewassen. Met de huidige marktprijzen zijn bio-energiegewassen niet concurrerend met conventionele agrarische gewassen op gronden die als ‘geschikt’ worden geclassificeerd. Op minder geschikte gronden hebben de intensief geteelde gewassen een laag rendement en behalen meerjarige gewassen een betere NCW dan

321

SHADES OF GREEN

conventionele gewasrotaties. De resultaten laten zien dat de minimale biomassaproductiekosten 5.4 €/GJ zijn voor Miscanthus en 9.7 €/GJ voor suikerbiet. Ethanolproductie van Miscanthus (24 €/GJ) is goedkoper dan van suikerbiet (27 €/GJ). Bioethanol productie van in Noord Nederland geproduceerde energiegewassen is echter niet concurrerend met benzine (12 €/GJ) onder huidige marktomstandigheden. Dit kan veranderen door een hogere olieprijs (in de berekeningen is een olieprijs van 62 $/barrel gehanteerd) en door meer kosteneffectieve bio-energieproductie middels bioraffinage en technologisch leren in de gehele productieketen. De analyse resulteert in een generieke methodologie om vanuit een economisch perspectief kansrijke locaties voor bioenergiegewassen te identificeren, rekening houdend met de ruimtelijke variatie in huidig landgebruik en biofysische factoren zoals bodemkwaliteit en beschikbaarheid van water. In Hoofdstuk 3 komen onderzoeksvragen II, III en IV aan bod. In dit hoofdstuk is de ruimtelijke variatie in potentiële milieu-impacts kwantitatief geanalyseerd. De teelt van suikerbieten en Miscanthus voor bioethanol productie in Noord Nederland is gebruikt als een casestudie. De milieu-impacts die zijn meegenomen in deze studie zijn: broeikasgasemissies (gedurende de levenscyclus en gerelateerd aan directe veranderingen in landgebruik), bodemkwaliteit, watergebruik en kwaliteit en biodiversiteit. Gebaseerd op een uitgebreide literatuurstudie zijn voor alle milieu-impacts geschikte indicatoren en methoden geselecteerd en zo nodig aangepast om ze ruimtelijk expliciet toe te kunnen passen. De ruimtelijke variatie in milieu-impacts is gerelateerd aan de ruimtelijke heterogeniteit van het fysieke milieu (bodem, water, klimaatlandgebruik, vegetatie) en is geanalyseerd met behulp van een Geografisch Informatie Systeem (GIS). De casestudie laat zien dat er grote ruimtelijke variaties zijn in de milieu-impacts van de introductie van bio-energiegewassen. In het algemeen resulteert de teelt van suikerbiet in negatieve milieueffecten, met name in de natte weidegebieden. In deze gebieden kunnen de -1 broeikasgasemissies door verandering in landgebruik oplopen tot 148 kg CO2-eq GJethanol -1 en het risico van erosie kan toenemen tot 9 ton grond ha . Daarnaast is er een groot risico op verlies van biodiversiteit in deze gebieden. Positieve effecten in deze gebieden zijn een -1 afname van 75 mg l in de NO3 concentratie in grondwater en een afname van 100 mm in de seizoensgebonden watertekorten. Wanneer akkerland wordt omgezet naar Miscanthus -1 kunnen de broeikasgasemissies tot 159 kg CO2-eq GJethanol worden gereduceerd, -1 -1 erosierisico kan worden verminderd met 4 ton ha , de NO3 concentratie kan met 53 mg l worden gereduceerd, alsmede kunnen positieve effecten op de biodiversiteit worden gerealiseerd. Daarentegen kunnen seizoensgebonden watertekorten toenemen tot 150 mm per jaar. Voor beide gewassen geldt dat in de westelijke natte weide gebieden de meeste negatieve effecten optreden. De gecombineerde resultaten laten zien dat er verschillende ‘trade-offs’ tussen milieu-impacts zijn: er zijn geen gebieden waar helemaal geen negatieve effecten optreden. De studie demonstreert een raamwerk dat het 322

Samenvatting en conclusies

mogelijk maakt om gebieden te identificeren waar mogelijke negatieve effecten van de teelt van energiegewassen optreden en gebieden waar de teelt van bio-energiegewassen netto nauwelijks of positieve milieu-impacts veroorzaken. Hoofdstuk 4 behandelt onderzoeksvraag I, II en IV door een model te ontwikkelen om de toekomstige beschikbaarheid van land voor bio-energiegewassen ruimtelijk expliciet en in de tijd te onderzoeken rekening houdend met het vermijden van competitie met andere landgebruikfuncties en derhalve het voorkomen van indirecte veranderingen in landgebruik. De landspecifieke factoren die bepalend zijn voor de ontwikkeling van andere landgebruiksfuncties zoals de productie van voedsel, veevoeder en materialen en de onzekerheid in deze factoren zijn geïdentificeerd en opgenomen in het model. Dit spatiotemporele, op PCRaster gebaseerde landgebruiksveranderingsmodel (PLUC) is gedemonstreerd aan de hand van de ontwikkelingen in beschikbaarheid van land voor bioenergiegewassen in Mozambique voor de periode 2005-2030. De ontwikkelingen in de belangrijkste factoren die bepalend zijn voor agrarisch landgebruik (i.e. vraag naar voedsel, dierlijke producten en materialen) is geanalyseerd op basis van de verwachte ontwikkelingen in populatiegroei, voedsel consumptie, bruto nationaal product en de mate van zelfvoorzieningszekerheid. Er zijn twee scenario’s ontwikkeld: een ‘Business-AsUsual’ (BAU) scenario en een progressief scenario. De allocatie van land is gebaseerd op een set geschiktheidfactoren specifiek voor elke landgebruiksklasse. De dynamiek in de veranderingen in landgebruik is voor elk jaar tot en met 2030 in kaart gebracht op een 2 resolutie van 1 km . In 2030 kan in het BAU scenario 7,7 Mha beschikbaar komen en in het progressieve scenario kan 16,4 Mha beschikbaar komen voor de teelt van bioenergiegewassen. Op basis van de Monte Carlo analyse, kan een 95% betrouwbaarheidsinterval van de hoeveelheid beschikbaar land en de ruimtelijke expliciete waarschijnlijkheid van de beschikbaarheid worden bepaald. De ‘bottum-up’ benadering, het aantal dynamische landgebruiksklassen, het diverse portfolio van bepalende factoren van veranderingen in landgebruik en van de allocatie van land, en de mogelijkheid om onzekerheden ruimtelijk expliciet te modeleren zorgt voor een aanzienlijk verbetering in de integrale modellering van de beschikbaarheid van land voor bio-energiegewassen. Daarnaast biedt het mogelijkheden om randvoorwaarden te verkennen voor een hoge inzetbaarheid van bio-energie terwijl indirecte veranderingen in landgebruik worden vermeden. Hoofdstuk 5 adresseert onderzoeksvragen II, III, en IV door te onderzoeken hoe het potentieel en de kosten van bio-energieproductie zich ruimtelijk door de tijd ontwikkelen. De analyse van de kosten en het potentieel is gebaseerd op de ontwikkelingen in de beschikbaarheid van land voor biomassa teelt, de geschiktheid van het land dat beschikbaar is en kan komen, de gedesaggregeerde kostenopbouw van de teelt van 323

SHADES OF GREEN

energiegewassen, de transportafstand van de biomassa naar de conversie-installatie, de conversiekosten, de transportafstand van de conversie-installatie naar de haven en de kosten voor internationaal transport. De productieketens van (getorrificeerde) pellets van eucalyptus en van ethanol uit suikerriet in Mozambique dienen als casestudie. De ontwikkelingen in de beschikbaarheid van land voor energiegewassen in Mozambique zijn gebaseerd op de resultaten van Hoofdstuk 4. De resultaten van Hoofdstuk 5 laten een grote spreiding zien in de kosten van de bio-energieketens als gevolg van de grote ruimtelijke variatie in de kosten van biomassa teelt en van primair en secundair transport. De meest veelbelovende gebieden voor de teelt van eucalyptus en suikerriet liggen verspreid in het centrale zuiden, het midden en het noordoosten van Mozambique met relatief gunstige agro-ecologische condities, voldoende potentiële beschikbaarheid van biomassa om te voldoen aan de capaciteit van de conversie-installatie en in de nabijheid van infrastructuur. In het progressieve scenario bedraagt het totale potentieel van eucalyptus pellet productie 3200 PJ in 2030, waarvan 2500 PJ onder een marktprijs van 8 €/GJ naar Europa kan worden geëxporteerd. Voor suikerriet ethanol is het totaal berekende potentieel 850 PJ in 2030, waarvan 500 PJ onder een prijsniveau van 30 €/GJ naar Europa kan worden geëxporteerd. De locatie van productie is de belangrijkste factor voor kosteneffectieve productie. Dit geldt in het bijzonder voor landen met een grote heterogeniteit in agro-ecologische condities en met een matig tot slecht ontwikkelde infrastructuur. Door verbeteringen aan weg- en spoor infrastructuur, kunnen de kosten voor logistiek gereduceerd worden en wordt bio-energieproductie op meer plaatsen economisch haalbaar. Deze studie demonstreert hoe de ontwikkeling in potentiëlen en kosten van bio-energie ruimtelijk expliciet en door de tijd kunnen worden gekwantificeerd. Omdat de milieu- en socio-economische impacts van bio-energie productie sterk afhangen van de biofysische en socio-economische condities van de productielocatie, is de ruimtelijk expliciete bepaling van bio-energie potentiëlen een belangrijke stap voor het ontwerpen, optimaliseren en het analyseren van de impacts van bio-energieketens. Hoofdstuk 6 behandelt onderzoeksvragen I, II, III en IV door te onderzoeken hoe bioenergie potentiëlen en de reductie van broeikasgasemissies in Oekraïne zich ruimtelijk expliciet en in de tijd kunnen ontwikkelen in de periode 2010-2030, rekening houdend met de ontwikkelingen en broeikasgasemissies van andere agrarische landgebruiksfuncties. De ontwikkelingen in de hoeveelheid land die nodig is voor voedselen veevoederproductie is op jaarlijkse basis ruimtelijk geanalyseerd met behulp van het PCRaster landgebruiksveranderingsmodel (PLUC). Het model is geschikt gemaakt voor Oekraïne door de dynamische landgebruiksklassen en de allocatiefactoren en hun eigenschappen aan te passen aan de Oekraïense situatie. Er zijn twee scenario’s voor de periode 2010-2030 geëvalueerd: een ‘Business as Usual’ (BAU) scenario, waarin de huidige 324

Samenvatting en conclusies

trends in landbouwproductiviteit worden gecontinueerd; en een progressief scenario, waarin het opbrengstniveau van Oekraïne convergeert met het niveau van de West Europese landen. In het progressieve scenario kan in 2030 32,1 Mha beschikbaar komen voor bio-energieproductie. De vrijgekomen landbouwgebieden liggen voornamelijk in het noorden, oosten en zuiden van Oekraïne. De geprojecteerde veranderingen in landgebruik dienen als input voor de spatiotemporele berekeningen van de broeikasgasbalans. De broeikasgasbalans bestaat uit de CO2, N2O and CH4 emissies door veranderingen in management en landgebruik evenals de vermeden emissies door het vervangen van fossiele brandstoffen door bioethanol uit tarwe en switchgrass. Het modeleren van de 2 broeikasgasemissies resulteert in ruimtelijk expliciete kaarten (1 km resolutie) van de individuele broeikasgasemissies voor elk jaar in de periode 2010-2030. Er wordt veel koolstof in de bodem vastgelegd wanneer verlaten landbouwgrond wordt gebruikt voor de teelt van switchgrass of voor herstel van natuurlijke vegetatie. De totale N2O emissies nemen toe wanneer verlaten landbouwgrond wordt gebuikt voor de teelt van tarwe en switchgrass, met name in de gebieden waar akkerbouw uitbreidt ten koste van gecombineerde akkerbouw-weide gebieden. De resultaten laten zien dat in het progressieve scenario een totale netto vermeden emissie van 0,8 GT CO2-eq voor tarwe en 3,8 GT CO2-eq voor switchgrass gehaald kan worden in 2030. De vermeden emissies van tarwe wordt mede verklaard door de groei van natuurlijke vegetatie op verlaten landbouwgronden die zijn uitgesloten voor de teelt van bio-energiegewassen. Wanneer er extra maatregelen worden getroffen voor de reductie van broeikasgasemissies in de landbouw (zoals verminderde grondbewerking, verhoogde toevoer van organische stof, en het gebruik van verbeterde meststoffen), kan de cumulatieve broeikasgasbalans zelfs oplopen tot -2,6 GT CO2-eq voor tarwe en -5,0 GT CO2-eq voor switchgrass in 2030. Wanneer het beschikbare land wordt gebruikt voor de regeneratie van natuurlijke vegetatie zal een aanzienlijke hoeveelheid koolstof worden geaccumuleerd in de vorm van biomassa en bodemkoolstof. Dit kan oplopen tot -4,3 GT CO2-eq in 2030. Deze koolstofvastlegging kan echter slechts eenmaal plaatsvinden in tegenstelling tot de reductie van broeikasgasemissies door bio-energiegewassen. Uitgaande van het progressieve scenario en een termijn tot 2100, wordt de totale cumulatieve broeikasgasemissiereductie geschat op ± 5,5 GT CO2 voor de regeneratie van natuurlijke vegetatie, 6 GT CO2-eq voor tarwe and 15 GT CO2-eq voor switchgrass. De spatiotemporele broeikasgasmodule in combinatie met het PLUC model maakt ruimtelijke expliciete en dynamische modellering van broeikasgasemissies als gevolg van veranderingen in landgebruik en -management door de implementatie van de teelt van bio-energiegewassen mogelijk.

325

SHADES OF GREEN

Op basis van de ontwikkelde methoden en de resultaten van Hoofdstuk 2 tot en met 6 worden hieronder antwoorden op de belangrijkste onderzoeksvragen en aanbevelingen voor beleid en onderzoek gegeven.

7.4

Belangrijkste bevindingen en conclusies I.

Hoe kan de potentiële beschikbaarheid van land voor bio-energiegewassen ruimtelijk expliciet en in de tijd worden bepaald, gegeven dat indirecte veranderingen in landgebruik moeten worden voorkomen en derhalve rekening houdend met de ontwikkeling in andere landgebruikfuncties?

Binnen dit onderzoek werd een nieuw landgebruiksveranderingsmodel (PLUC) ontwikkeld om de ontwikkelingen in de beschikbaarheid van land voor bio-energiegewassen op een 2 ruimtelijk gedetailleerd niveau (1 km ) te onderzoeken, rekening houdend met de dynamiek en de onzekerheid van de belangrijkste factoren die bepalend zijn voor veranderingen in landgebruik. De belangrijkste factoren die de vraag naar landbouwproducten in een regio beïnvloeden zijn de ontwikkelingen in populatie, bruto nationaal product, consumptie per hoofd van de bevolking en de mate van zelfvoorzieningszekerheid. De totale hoeveelheid land die nodig is om in de vraag naar voedsel, hout en dierlijke producten te kunnen voorzien is direct afhankelijk van de agro-ecologische geschiktheid van de locatie voor een specifieke landgebruiksklasse (e.g. akkerland, weiland, bos) en van de efficiëntie van de agrarische sector. Gegeven de onzekerheden in hoe de genoemde factoren zich in de toekomst ontwikkelen, wordt gebruikt gemaakt van scenario’s. De allocatie van land is gebaseerd op de geschiktheid van de locatie voor een specifieke. De geschiktheid wordt bepaald door een combinatie van ruimtelijke expliciete allocatiefactoren die betrekking hebben tot de biofysische eigenschappen, de bereikbaarheid, de conversie-elasticiteiten en het landgebruik in de omgeving. Voor elke allocatiefactor moet de richting, de aard, en de mate van correlatie met het landgebruik worden bepaald. De allocatiefactoren, hun eigenschappen alsook hun relatieve gewicht zijn specifiek voor de regio- en voor de landgebruiksklasse. Gebieden die niet geschikt zijn (e.g. steile hellingen) of niet mogen worden omgezet (e.g. beschermde gebieden), worden uitgesloten voor agrarische productie. Het startpunt van de allocatie van landgebruik is de kaart van het huidige landgebruik gekalibreerd met de statistieken van de vraag naar voedsel, veevoeder en materialen, de landbouwproductiviteit, en de gemiddelde agro-ecologische geschiktheid per landgebruiksklassen. Land wordt gealloceerd aan landgebruiksklassen in tijdstappen van een jaar. De totale allocatie van land in één tijdstap is afgerond wanneer alle landgebruiksklassen zijn gealloceerd en de gesommeerde productie van de

326

Samenvatting en conclusies

landgebruiksklassen aan de totale vraag van dat jaar voldoen. Het model bevat een terugkoppeling: de landgebruikkaart resulterend van de landgebruikallocatie in tijdstap t dient als input voor de landallocatie in tijdstap t+1. De output van dit type modeleren van landgebruikverandering zijn kaarten van landgebruik en de beschikbaarheid van land voor 2 bio-energiegewassen op een ruimtelijk gedetailleerd niveau (1 km ) voor elk jaar binnen de gekozen tijdsperiode. Het grote voordeel van dit modelraamwerk is de mogelijkheid om stochastische invoergegevens te gebruiken. Dit maakt het mogelijk om spatiotemporele Monte Carlo (MC) analyses uit te voeren, waarmee de voortplanting van onzekerheden kan worden geëvalueerd. Met PLUC kunnen tijdseries (e.g. gewasvraag en agrarische productiviteit), ruimtelijke invoerparameters (e.g. populatiedichtheid en agro-ecologische geschiktheid), en eigenschappen van geschiktheidsfactoren (e.g. de maximale afstand met effect in de afstand tot infrastructuur) stochastisch worden gemodelleerd. Dit resulteert in kaarten die 2 de waarschijnlijkheid van de beschikbaarheid van een specifieke locatie (gridcel, 1 km ) in een specifieke tijdstap (jaar) weergeven. Het landgebruiksveranderingsmodel dat is ontwikkeld in deze studie is een geavanceerd manier om mogelijke toekomstige dynamiek in landgebruik en landbeschikbaarheid voor bio-energiegewassen te analyseren. Het gebruik van een scenariobenadering voor de belangrijkste bepalende factoren van landgebruiksveranderingen en het toepassen van een ‘food-first’ paradigma, maakt het mogelijk te onderzoeken wat het biomassapotentieel is dat gerealiseerd kan worden zonder dat er competitie met voedsel en veevoeder productie ontstaat. Daarbij kan bepaald worden wat de benodigde randvoorwaarden zijn om deze potentiëlen te realiseren. De ‘bottom-up’ benadering, het aantal dynamische landgebruiksklassen, het diverse portfolio van bepalende factoren van landgebruiksverandering en allocatiefactoren en de mogelijkheid om onzekerheid ruimtelijk te modeleren, betekent een stap voorwaarts in het bepalen van beschikbaarheid van land voor energiegewassen. Bovendien kan deze benadering bijdragen aan het identificeren van de randvoorwaarden voor het realiseren van deze potentiëlen. Omdat de biomassa opbrengsten, productiekosten, logistiek en milieuimpacts sterk afhangen van de locatiespecifieke biofysische condities (e.g. agroecologische geschiktheid, beschikbaarheid van infrastructuur, bodem eigenschappen, klimaat condities etc.); is de ruimtelijke expliciete bepaling van landbeschikbaarheid voor bio-energiegewassen een belangrijke basis voor het gedetailleerd analyseren van milieuen socio-economische impacts van biomassaproductie.

327

SHADES OF GREEN

Het model is nu toegesneden op, en gedemonstreerd voor Mozambique en Oekraïne. De modelstructuur is echter flexibel en kan worden aangepast voor andere landen en regio’s wanneer invoerdata, allocatie regels en eigenschappen van allocatiefactoren bekend zijn. II.

Hoe kunnen de economische haalbaarheid (de locatiespecifieke concurrentie met andere agrarisch grondgebruiken, de kosten van de biomassa productie en de logistiek van de biomassa keten) en de milieuimpacts (broeikasgasemissies, bodem, water en biodiversiteit) van bioenergie productie ruimtelijk en tijdsafhankelijk worden bepaald?

In het begin van dit onderzoek zijn methoden ontwikkeld om de economische- en milieuprestaties van bio-energieproductie ruimtelijk expliciet te onderzoeken rekening houdend met de ruimtelijke variabiliteit in de biofysische context. In een later stadium is een spatiotemporeel model ontwikkeld om de economische haalbaarheid en de broeikasgasemissies te onderzoeken, rekening houdend met zowel ruimtelijke variatie als de ontwikkelingen in de tijd. De economische haalbaarheid van bio-energieproductie is afhankelijk van de rentabiliteit van de teelt van bio-energie gewassen ten opzichte van ander landgebruik en het concurrentievermogen van bio-energieproductie ten opzichte van het referentie energiesysteem. De drie belangrijke kostenfactoren van bio-energieproductie zijn: de kosten van biomassa productie, de kosten van de logistieke keten, en de kosten voor conversie. De biomassa productiekosten kunnen worden bepaald door de netto contante Waarde (NCW) te berekenen van alle kosten (land, arbeid, machines, agrarische inputs) en de biomassa opbrengst gedurende de levensduur van de biomassa plantage. De ruimtelijke variatie in de opbrengst en gerelateerde kosten en het concurrentievermogen ten opzichte van ander agrarische landgebruik kan worden bepaald door de rentabiliteitberekeningen te combineren met de landgebruikkaart en de gewasspecifieke agro-ecologische geschiktheidkaarten. De ruimtelijke variatie in de kosten van biomassalogistiek kan worden berekend aan de hand van de data van de schaal van de conversie-installatie en de ruimtelijke informatie over landbeschikbaarheid, opbrengstniveaus, en de beschikbaarheid en de kwaliteit van infrastructuur. De conversiekosten omvatten de investeringskosten, de exploitatie- en onderhoudskosten en kosten voor energie en andere inputs. De kosten van biomassaproductie en conversie veranderen in de tijd door technologisch leren. Door de projecties van technologisch leren te integreren in de kostenberekeningen en te koppelen met de spatiotemporele modellering van landgebruiksveranderingen, kunnen de economisch prestaties van bioenergieketens ruimtelijk expliciet en in de tijd bepaald worden.

328

Samenvatting en conclusies

De selectie van milieu-impacts van bio-energieproductie die in dit proefschrift geanalyseerd worden, is gebaseerd op de criteria die door verschillende nationale en internationale initiatieven op het gebied van duurzaamheid van bio-energieproductie zijn geformuleerd (EC 2009a; NEN 2009; RSB 2010). De milieu-impacts die zijn meegenomen in dit proefschrift zijn de broeikasgasbalans (van de levenscyclus en gerelateerd aan landgebruiksveranderingen) en de impacts op water, bodem en biodiversiteit. De milieuimpacts zijn kwantitatief en ruimtelijk expliciet beoordeeld door methoden te ontwikkelen of bestaande methoden aan te passen voor ruimtelijk expliciet en gedetailleerde analyse. De impacts ten aanzien van water zijn geanalyseerd door een eenvoudige waterbalans te berekenen en gebruik te maken van gedetailleerde ruimtelijk expliciete data van veranderingen in landgebruik, de effectieve neerslag en de potentiële evapotranspiratie. De impacts op de bodem zijn gekwantificeerd aan de hand van de winderosie vergelijking (WEQ), rekening houdend met ruimtelijke variatie in veranderingen in landgebruik, vegetatie karakteristieken, bodem eigenschappen, en wind-, neerslag- en temperatuurpatronen. Het effect op biodiversiteit is onderzocht door de ‘Mean Species Abundance’ (MSA) en de ‘High Nature Value’ (HNV) indicatoren toe te passen en gebruik te maken van de ruimtelijke informatie over landgebruik, management en de distributie van natuurgebieden en bedreigde soorten. In de eerste hoofdstukken van dit proefschrift werd het Miterra model (Lesschen 2008; Velthof et al. 2009) gebruikt om de broeikasgasemissies en impacts op waterkwaliteit gerelateerd aan veranderingen in landgebruik en -management te kwantificeren. Dit deterministische, statische en op vector informatie gebaseerde model, simuleert de stikstof en fosfaat balans, NH3, N2O, NOx, and CH4 emissies, uitspoeling van stikstof, NO3 concentratie in grondwater en veranderingen in bodemkoolstof door veranderingen in landgebruik en -management. In een later studium is een nieuw model ontwikkeld om de 2 broeikasgasemissies dynamisch en op een ruimtelijk gedetailleerde niveau (1 km ) te berekenen. De berekeningen van de stikstof routes zijn primair gebaseerd op de methode ontwikkeld door Lesschen (2008) en Velthof et al. (2009) maar is aangepast voor rastergebaseerde en dynamische berekeningen. Daarnaast zijn in plaats van de standaard N2O-emissiefactor van de IPCC, de N2O-emissiefactoren specifiek voor stikstofbron, bodem en landgebruik ontwikkelt door Lesschen et al. (2011) geïmplementeerd in het model en zijn de uitspoeling en afvoer factoren aangepast. De berekeningen van de koolstofemissies zijn gebaseerd op de methode van de IPCC (2006) en dynamisch gemodelleerd. Door de integratie van de module van de broeikasgasemissies en het spatiotemporele landgebruiksveranderingmodel kunnen de broeikasgasemissies ruimtelijk expliciet en dynamisch worden berekend. Dit nieuwe model is in staat rekening te houden met de ruimtelijke variatie in landgebruik, opbrengst niveaus, bodemkenmerken, klimaat, helling, en temporele variatie in het landgebruik, management, en opbrengst niveaus. Het 329

SHADES OF GREEN

model maakt het mogelijk de beste gebieden voor intensivering van de landbouw en de meest geschikte gebieden voor de productie van bio-energiegewassen te identificeren. Bovendien kan met het model onderzocht worden onder welke voorwaarden de broeikasgasbalans van het totale landgebruik kan worden geoptimaliseerd terwijl de productie van voldoende voedsel en veevoeder gewaarborgd blijft. De ontwikkelde methoden en modellen bieden een aanpak om ex ante gebieden te identificeren waar de toepassing van bio-energieproductie economisch aantrekkelijk is of kan worden en waar weinig negatieve of positieve milieueffecten optreden. De vergelijking van de ruimtelijke variatie in de verschillende milieueffecten en in de economische haalbaarheid maakt evaluatie van de ‘trade-offs’ tussen milieueffecten en tussen milieu- en economische prestaties mogelijk. De integratie van deze modellen en methoden maakt het mogelijk de ‘go’ en ‘no-go’ gebieden voor de productie van bioenergieproductie te identificeren vanuit milieu- en economisch oogpunt, rekening houdend met het vermijden van concurrentie om land. III.

Wat zijn de potentiëlen, de economische prestaties, en de milieu-impacts van bio-energieproductie in verschillende regio’s in de wereld?

Wanneer indirecte landgebruikveranderingen (iLUC) moeten worden vermeden is de hoeveelheid bio-energie die kan worden geproduceerd afhankelijk van de ontwikkelingen in de vraag naar andere landbouwproducten, het tempo van de intensivering van de agrarische sector, en de geschiktheid van het land dat beschikbaar komt voor de teelt van energiegewassen. In Nederland neemt de bevolking langzaam toe en het wordt verwacht dat de voedselconsumptie per hoofd van de bevolking stabiel blijft op het huidige niveau. Nederland heeft een van de meest efficiënte en technologisch geavanceerde agrarische sectoren van Europa (de Wit et al. 2011a). Daarom zijn de verschillen tussen de gerealiseerde en de maximaal haalbare opbrengsten en de daarmee gerelateerde mogelijkheden voor efficiëntieverbeteringen relatief beperkt (zeker in vergelijking met Oekraïne en Mozambique). Dit resulteert in een relatief lage potentiële beschikbaarheid van land voor bio-energiegewassen, variërend van 41 tot 51 kha in 2030 in de drie noordelijke provincies (de Wit and Faaij 2010; Fischer et al. 2010a; Fischer et al. 2010b). Uitgaande van een regionaal gemiddelde opbrengst van Miscanthus en suikerbieten kan er in Groningen, Friesland en Drenthe gezamenlijk 6 tot 9 PJ ethanol worden geproduceerd in 2030. Omdat niet onderzocht is waar het land beschikbaar komt, kan er geen precieze schatting worden gemaakt van het productiepotentieel van bioenergiegewassen. Uitgaande van de laagste biomassa productiekosten bedragen de

330

Samenvatting en conclusies

kosten voor de productie van bioethanol 24 €/GJ voor Miscanthus en 27 €/GJ voor suikerbiet ethanol (uitgaande van de kosten en efficiënties van de conversietechnologieën die op kortere termijn beschikbaar kunnen zijn). In Mozambique zal de vraag naar agrarische producten aanzienlijk toenemen als gevolg van een sterk groeiende bevolking en een toename van de voedselconsumptie per capita. Anderzijds biedt de huidige lage landbouwproductiviteit veel ruimte voor verbetering. Wanneer er van het ‘Business as Usual’ (BAU) scenario wordt uitgegaan waarin de huidige trends in de landbouwproductiviteit worden voortgezet, zal in 2030 7,7 Mha beschikbaar kunnen komen voor bio-energiegewassen. In het progressieve scenario, waarin de landbouwproductiviteit aanzienlijk verhoogd wordt, kan 16,4 Mha beschikbaar komen. Als de beschikbare grond wordt gebruikt voor eucalyptusproductie kan in 2030 1340 PJ getorrificeerde pellets worden geproduceerd in het BAU-scenario en in 3200 PJ in het progressieve scenario. Wanneer de beschikbare grond wordt gebruikt voor het verbouwen van suikerriet voor de productie van ethanol kan in 2030 350 PJethanol in het BAU en 850 PJethanol in het progressieve scenario worden geproduceerd. De ondergrens van de productiekosten bedraagt 5 €/GJ voor getorrificeerde pellets en 14 €/GJ voor ethanol uit suikerriet. De gebieden met de laagste productiekosten zijn de zones met relatief gunstige agro-ecologische condities, voldoende beschikbaarheid van land om te voldoen aan de capaciteit van de conversie-installatie en in de nabijheid van infrastructuur en havens. Wanneer weg- en spoorinfrastructuur worden verbeterd kunnen de kosten van logistiek worden gereduceerd en kunnen in de toekomst meer gebieden rendabel worden voor bio-energie productie. In Oekraïne zal de vraag naar agrarische producten tot 2030 naar verwachting licht toenemen. Ofschoon de populatie afneemt en voedselconsumptie per hoofd van de bevolking stabiliseert, wordt verwacht dat de totale productie toeneemt omdat Oekraïne haar exportvolume naar verwachting zal vergroten. Ondanks de gunstige agro-ecologische condities is de productiviteit van de agrarische sector laag, wat veel mogelijkheden biedt voor verbetering. In het BAU scenario, waarin weinig tot geen verbeteringen in de landbouw worden gerealiseerd, komt een beperkte hoeveelheid land (0,03 Mha) beschikbaar voor bio-energie productie. Echter, in het progressieve scenario zal in 2030 32,1 Mha beschikbaar kunnen komen voor de teelt van energiegewassen. Wanneer het beschikbare land wordt gebruikt voor de teelt van switchgrass voor ethanol kan 3 PJethanol in het BAU en 2230 PJethanol in het progressieve scenario worden geproduceerd in 2030. Wanneer tarwe wordt geteeld op het netto beschikbare agrarische land, kan 1 tot 2370 PJethanol worden geproduceerd in 2030. De productie kosten van bioethanol variëren tussen 9 tot 11 €/GJ en zijn berekend op basis van de minimale kosten van biomassaproductie in Oekraïne gebruik makend van gegevens uit de Wit en Faaij (2010). 331

SHADES OF GREEN

Omdat het landgebruik in Oekraïne wordt gedomineerd door landbouw (79%), en het merendeel van het overige land uit bos (15%) of statische vormen van landgebruik bestaat (5%, zoals natuurgebieden, stedelijk gebied), is er weinig land beschikbaar voor de teelt van energie gewassen, tenzij de hoeveelheid land die nodig is voor voedsel en vee afneemt door middel van efficiëntere agrarische productie. In Mozambique is er op dit moment relatief veel land beschikbaar voor de teelt van bio-energiegewassen (9 Mha). Naast het huidige land dat in gebruik is voor landbouw (20%), bos (60%) en statische vormen van landgebruik (9%), zou ongeveer 10% van het land gebruikt kunnen worden voor bio-energiegewassen. De beschikbaarheid van land voor bio-energiegewassen en bos neemt echter af wanneer de hoeveelheid land die nodig is voedsel en voeder productie groeit als gevolg van de toenemende vraag en tegelijkertijd weinig tot geen verbeteringen in de landbouwefficiëntie worden gerealiseerd. Vanwege de grote verschillen in biofysische karakteristieken, sociaaleconomische condities, historische ontwikkelingen, stadium van ontwikkeling en beleidspecifieke context, zijn er grote verschillen in de factoren die de dynamiek in landgebruik in Mozambique en Oekraïne bepalen. Vanwege de grote afhankelijkheid van zelfvoorzienende landbouw, het gebrek aan infrastructuur en de afhankelijkheid van lokale markten, wordt verwacht dat landbouwgrond in Mozambique zich zal concentreren in en rond de dichter bevolkte gebieden, nabij grote steden en dicht bij het wegennet. In Oekraïne is een minder groot deel van de bevolking direct afhankelijk van de landbouw, is er een relatief fijnmazig wegennet beschikbaar, en is er een geleidelijke verschuiving zelfvoorzienende landbouw naar grotere, commerciële landbouwbedrijven aan de gang. Hier is de locatie van landbouwgrond vooral afhankelijk van de agro-ecologische geschiktheid en in mindere mate gerelateerd aan de bevolkingsdichtheid of de toegankelijkheid van een gebied. Vanwege deze grote verschillen zijn de parameters in het PLUC model (bepalende factoren van veranderingen in landgebruik, allocatiefactoren hun kenmerken en hun relatieve gewicht) aangepast. Uit de analyse van de kostenontwikkeling van de bio-energieketen blijkt dat de kosten aanzienlijk kunnen worden gereduceerd door middel van het verbeteren van de biomassaopbrengsten, optimalisatie van de logistieke keten en verbeteringen in de voorbehandelings- en conversietechnologieën. De geïntegreerde analyse van de milieueffecten in het noorden van Nederland laat een sterke ruimtelijke variatie in impacts zien. Er zijn geen gebieden waar exclusief positieve effecten optreden wanneer bio-energiegewassen worden geïntroduceerd, en er zijn ‘trade-offs’ tussen de impacts. In het algemeen resulteert de teelt van suikerbiet in 332

Samenvatting en conclusies

negatieve milieueffecten, met name in de natte weidegebieden. In deze gebieden kunnen de broeikasgasemissies door verandering in landgebruik oplopen tot 148 kg CO2-eq -1 GJethanol en het risico van erosie kan toenemen tot 9 ton grond per hectare per jaar. Daarnaast is er een groot risico op verlies van biodiversiteit in deze gebieden. Positieve -1 effecten in deze gebieden zijn een afname van 75 mg l in de NO3 concentratie in grondwater en een afname van 100 mm in de seizoensgebonden watertekorten. Wanneer akkerland wordt omgezet naar Miscanthus kunnen de broeikasgasemissies tot 159 kg CO2-1 -1 eq GJethanol worden gereduceerd, erosierisico kan worden verminderd met 4 ton ha , de -1 NO3 concentratie kan met 53 mg l worden gereduceerd, alsmede kunnen positieve effecten op de biodiversiteit worden gerealiseerd. Daarentegen kunnen seizoensgebonden watertekorten toenemen tot 150 mm per jaar. De ruimtelijke patronen van de effecten zijn vooral gekoppeld aan het ruimtelijk patroon van het huidige landgebruik, dat wil zeggen dat de omvang van de milieu-impact grotendeels wordt bepaald door het type landgebruik dat wordt omgezet naar energiegewassen. Uit de geïntegreerde analyse van de broeikasgasbalans van de teelt van bio-energiegewassen in combinatie met de intensivering van de agrarische sector blijkt dat een hoog reductiepotentieel van broeikasgasemissies kan worden bereikt wanneer agrarisch landgebruik wordt geïntensiveerd en biomassa teelt wordt geïntroduceerd in de verlaten landbouwgebieden. Hoge emissiereducties zijn alleen mogelijk wanneer de gehele agrarische sector duurzamer wordt beheerd. Globale vergelijking van een aantal kwantitatieve resultaten In Tabel 7.2 wordt een overzicht van de potentiëlen, kosten en milieu-impacts van bioenergie productie gegeven voor de verschillende geografische regio’s die in dit proefschrift een rol spelen. In Mozambique en Oekraïne kunnen in de progressieve scenario's aanzienlijke hoeveelheden land beschikbaar komen voor de teelt van energiegewassen zonder dat er conflicten ontstaan met andere landgebruiksvormen. Dit kan mogelijk resulteren in grote emissiereducties van broeikasgassen. Echter, de lage en dalende potentiële beschikbaarheid van land in het BAU scenario wijst op een toenemende concurrentie om land in de toekomst. Daarom moet worden benadrukt dat een grootschalige duurzame bio-energie sector alleen kan worden gerealiseerd wanneer er tegelijkertijd een meer productieve en duurzame agrarische sector wordt ontwikkeld. Dit is nadrukkelijk een trendbreuk, wat voor Mozambique een verschuiving van zelfvoorzienings- naar commerciële landbouw, en van pastorale naar gemengde veeteelt systemen inhoud. Dit vereist veranderingen in het agrarisch management (vooral in het gebruik van kunstmest en verbeterde zaden), de ontwikkeling van regionale en /of (inter-) nationale markten, verbeterde logistiek, training en beter beheer en bestuur van de agrarische sector. Voor 333

SHADES OF GREEN

Oekraïne zijn landhervormingen cruciaal omdat die de transitie naar commerciële landbouw moeten faciliteren, normen en standaarden moeten aangepast worden aan internationale eisen voor landbouwproducten, en economische hervormingen zijn nodig voor het creëren van banen buiten de landbouw en een betere toegang van agrarische ondernemers tot kapitaal en hoogwaardige technologie en inputs is vereist. Tabel 7.2: Overzicht van de potentiëlen, kosten en milieu-impacts van bio-energie voor de verschillende geografische regio’s die in dit proefschrift een rol spelen. Eenheid Nederland Mozambique Oekraïne Land beschikbaarheid a Mha 0,04 – 0,05 7,7 -16,4 0,03 – 32,1 Bioethanol productie potentieel 1st gen ethanol b PJ 6-7 350 - 850 1 - 2370 Bio-energie productie potentieel van hout- en grasachtige gewassen c PJ 7-9 1340 - 3200 3 - 2230 Ethanol kosten 1e gen ethanol d €/ GJ 27 - << 14 - << ~11 - << Bioenergy kosten 2e gen ethanol / getorrificeerde €/ GJ 24 - << 5 - << ~9 - << pellets d Milieu impact 1e gen bioenergy gewassen e -/+ Milieu impacts hout- en grasachtige gewassen e +/+ a Voor Nederland zijn alleen de gebieden inbegrepen die momenteel in gebruik zijn als landbouwgrond. De beschikbaarheid is gebaseerd op de bevindingen van de Wit en Faaij (2010). De beschikbare grond in Mozambique en Oekraïne omvat ook andere gebieden voornamelijk bestaande uit grasland en struikgewas (bos, beschermde gebieden, steile hellingen, etc zijn uitgesloten). b Betreft ethanol van suikerbieten in Nederland, van suikerriet in Mozambique en van tarwe in Oekraïne. c Betreft ethanol van Miscanthus in Nederland, getorrificeerde houtpellets van eucalyptus in Mozambique, en ethanol van switchgrass in Oekraïne. d De productiekosten van ethanol in Oekraïne zijn gebaseerd op de minimale biomassakosten afgeleid van de Wit en Faaij (2010). De kostencijfers gegeven voor Mozambique omvatten transport van de conversie-installatie naar de haven, opslag en verscheping. De kostencijfers van ethanolproductie in Nederland en Oekraïne zijn de kosten tot en met conversie en bevatten dus niet de kosten van transport naar de haven en internationale verscheping. e De milieueffecten die meegenomen zijn in Nederland zijn broeikasgasemissies, en de impact op water, bodem en biodiversiteit. In Oekraïne zijn de broeikasgasemissies onderzocht. Voor Mozambique zijn in dit proefschrift geen milieueffecten van bio-energieproductie onderzocht.

IV.

Welke betrouwbaarheid kan worden verkregen met behulp van de beschikbare gegevens en de in deze studie ontwikkelde methoden?

In dit proefschrift is een spatiotemporele analyse het productiepotentieel, de economische prestaties en de milieueffecten van bio-energie productie gemaakt. Deze ruimtelijke expliciete en ex ante analyses gaan gepaard met tal van onzekerheden. Sturende factoren van landgebruiksveranderingen De belangrijkste factoren die bepalend zijn voor veranderingen in landgebruik in verschillende situaties zijn inherent onzeker. Bovendien zijn de factoren die bepalend zijn

334

Samenvatting en conclusies

voor de locatie van landgebruiksveranderingen (zoals de agro-ecologische geschiktheid, de toegankelijkheid, de conversie elasticiteit, en het landgebruik in de omgeving) en de richting, het type, de mate van correlatie, en het relatieve gewicht lastig te identificeren en kunnen ze veranderen in de tijd. Daarom is validatie en kalibratie van het ruimtelijke gedetailleerde landgebruiksverandering model cruciaal. Longitudinale evaluatie door het monitoren van de veranderingen in zowel de bepalende factoren van landgebruik als in de landgebruikspatronen van verschillende landen, kan meer inzicht geven in de diverse relevante correlaties. Ontwikkeling in kosten en de economische haalbaarheid van bio-energieproductie Op de lange termijn zullen fossiele brandstoffen naar verwachting duurder worden wat bij zou kunnen dragen aan de haalbaarheid van bio-energieproductie. De afgelopen jaren zijn de biomassa productiekosten beïnvloed door de gestegen kosten van materieel, diesel, kunstmest en landbouwchemicaliën. Het is waarschijnlijk dat deze kosten blijven stijgen in lijn met de historische mondiale trends. Daarnaast zijn de productiekosten in landen als Mozambique en Oekraïne op dit moment relatief laag vanwege de lage kosten van met name arbeid en grond. Deze kosten zullen naar verwachting toenemen als gevolg van een verhoogde druk op land en door economische ontwikkeling. Anderzijds wordt verwacht dat door efficiënter management (bijvoorbeeld door verbeterde teelt, hogere opbrengsten, etc.) deze kostenstijgingen kunnen worden gecompenseerd. Competitie voor land Omdat de teelt van bio-energiegewassen niet is meegenomen als een dynamische landgebruiksklasse, en omdat een vaste allocatievolgorde van de andere landgebruiksklassen is toegepast, is de concurrentie om land of indirecte veranderingen in landgebruik niet expliciet gemodelleerd. In de praktijk is het echter waarschijnlijk dat de teelt bio-energiegewassen zal concurreren met andere landgebruikfuncties in de meest geschikte gebieden. Als de concurrentie tussen de bio-energiegewassen en andere vormen van landgebruik moet worden gemodelleerd, zal de implementatie van bioenergiegewassen gerelateerd moeten zijn aan een verwachte vraag (bijv. Nationale doelstellingen voor duurzame energie). Om de concurrentie tussen verschillende vormen van landgebruik goed te kunnen modelleren, is veel informatie nodig over ontwikkelingen in de markt, prijselasticiteit en beleid. Het belangrijkste uitgangspunt in de analyse van de beschikbaarheid van land was echter dat de competitie om land moet worden voorkomen, en daarom is al het land dat (potentieel) in gebruik is voor andere landgebruiksfuncties uitgesloten.

335

SHADES OF GREEN

Data beschikbaarheid en kwaliteit Ruimtelijk expliciete analyses van veranderingen in landgebruik, de milieu-impacts en de economische prestaties van bio-energieketens vereist grote hoeveelheden (ruimtelijke) data. De beschikbaarheid en kwaliteit van (ruimtelijke-) data zijn begrensd en er zijn inconsistenties tussen de verschillende bronnen van statistische data, tussen ruimtelijke datasets en tussen de statistische en ruimtelijke data. Daarbij kennen verscheidene ruimtelijke datasets ernstige kwaliteitsproblemen in termen van resolutie, classificatie, consistentie en documentatie. In de analyses zijn gegevens van meerdere bronnen gecombineerd, die niet noodzakelijkerwijs consistent zijn met elkaar. Voorbeelden hiervan zijn de combinaties van de data van agro-ecologische geschiktheid met de data van SOC, C: N ratio en klimaat. Voor een grotere nauwkeurigheid zijn betere ruimtelijke gegevens van de belangrijkste modelparameters zoals het huidige landgebruik, de bodemkenmerken (sediment, SOC, C:N ratio, grondwater niveau), het gebruik en de conditie van weilanden, de agro-ecologische geschiktheid voor verschillende vormen van landgebruik vereist. Hetzelfde geldt voor verbeteringen in de consistentie van de statistische gegevens van landgebruik, opbrengst niveaus, agrarische productie, veehouderij, etc. Uniformiteit - heterogeniteit In de analyse van de potentiëlen, kosten en milieu-impacts van biomassaproductie, is uitgegaan van nationale uniformiteit in de landgebruiksklassen, (ontwikkelingen in) agronomisch management en sturende factoren voor veranderingen in landgebruik. In de praktijk is er echter grote ruimtelijke variatie in deze parameters. Zo wordt bijvoorbeeld aangenomen dat de landgebruiksklasse ‘akkerland’ uit een gewogen sommering van alle geproduceerde gewassen bestaat, terwijl in de praktijk de samenstelling van gewasrotaties zal verschillen voor verschillende gebieden en locaties. Ook wordt aangenomen dat de landbouwpraktijken en de adoptie van verbeterde werkwijzen en daarmee samenhangende efficiëntie verbeteringen uniform zijn voor alle verschillende soorten producenten en voor alle agro-ecologische geschiktheidklassen. Alleen de opbrengst gerelateerde parameters zoals de kunstmest gift, opbrengst en oogst worden gevarieerd in relatie tot de agro-ecologische geschiktheid. In de praktijk zijn er op dit moment echter grote verschillen in de agrarische wekwijzen en in de ruimtelijke spreiding, bijvoorbeeld tussen kleinere en grotere producenten. Relaties tussen impacts De berekeningen van de milieu-impacts zijn gebaseerd op een breed scala van input parameters. Aangezien alle effecten gerelateerd zijn aan de functionaliteit van ecosystemen zijn ze sterk met elkaar verbonden. Het is echter gecompliceerd om causaliteit aan te tonen en relaties te kwantificeren.

336

Samenvatting en conclusies

Milieu-impacts kunnen beter gemodelleerd worden wanneer de modellering meer procesgebaseerd is in plaats van gebruik te maken van standaardwaarden en –factoren. Daarnaast is de beoordeling van de ernst van de veroorzaakte effecten complex, aangezien zowel de dosis-respons relaties als de drempelwaarden voor werkelijke schade locatie, tijd en schaalafhankelijk kunnen zijn. Analyse van onzekerheden De in dit proefschrift gedemonstreerde aanpak biedt nieuwe mogelijkheden voor het omgaan met onzekerheden binnen de spatiotemporele modellering van de potentiëlen, kosten en milieueffecten van bio-energieproductie. De gevoeligheidsanalyses laten zien in welke mate de onzekerheden in invoergegevens resulteren in veranderingen in de resultaten. De scenariobenadering maakt het mogelijk uiteenlopende veronderstellingen over de belangrijkste sturende factoren van veranderingen in landgebruik te modelleren. Het PLUC model kan omgaan met onzekerheden door middel van Monte Carlo analyses van stochastische inputparameters. De ontwikkeling van de potentiële veranderingen in 2 landgebruik, kosten en milieu-impacts zijn berekend op een 1 km resolutie, maar de hierboven genoemde factoren beperken de nauwkeurigheid van de resultaten. De huidig beschikbare data worden echter voldoende nauwkeurig beschouwd om patronen en ‘hotspots’ van veranderingen in landgebruik, economische haalbaarheid en milieu-impacts te kunnen onderscheiden. Daarom kunnen de resultaten worden gebruikt voor een eerste screening om 'go' en 'no-go' gebieden te identificeren. Afhankelijk van kwaliteitseisen, zijn voor toepassing van de gepresenteerde benadering voor de monitoring en certificering van biomassa productie betere data en een fijnere modelresolutie gewenst.

7.5

Aanbevelingen voor verder onderzoek •



Voor de analyse van de dynamiek van de competitie voor land zou de allocatie van landgebruik ook gebaseerd moeten zijn op de relatieve economische prestaties van de alternatieve vormen van landgebruik. Om rekening te houden met de ontwikkeling in prijzen, vraag en aanbod van biobrandstoffen, fossiele brandstoffen en landbouwproducten, is koppeling van landgebruiksveranderingsmodellen met algemene en / of partiële evenwichtsmodellen nodig. Longitudinale evaluatie door het monitoren van de ontwikkelingen in zowel de bepalende factoren als de patronen van veranderingen in landgebruik kan het inzicht in de correlatie tussen de sturende factoren en de veranderingen in landgebruik verbeteren en zorgen voor betrouwbaardere projecties van mogelijk toekomstig landgebruik.

337

SHADES OF GREEN











Meer kennis op het gebied van de onderlinge relaties tussen verschillende milieuimpacts is vereist. Daarnaast vergt de differentiatie van de dosis-respons relaties en drempelwaarden voor werkelijke schade in voor variaties in biofysische context, tijd en schaal verder onderzoek. Effecten op water zouden beoordeeld moeten worden op het niveau van een stroomgebied teneinde alle relevante mechanismen mee te nemen die bepalen of waterbalans in een gebied al dan niet teveel wordt verstoord. Tot slot is er meer validatie van en samenhang tussen de methoden en indicatoren nodig om de effecten van veranderingen in landgebruik op de biodiversiteit op verschillende ruimtelijke niveaus te kwantificeren en te monitoren. Sociaal-economische effecten van bio-energie productie zijn vaak sterk gerelateerd aan landgebruik, economische prestaties en milieu-impacts. Deze effecten moeten in combinatie worden geëvalueerd om de meest geschikte gebieden en wijze van productie te identificeren en om meer complexe indicatoren, zoals voedselzekerheid, te kwantificeren. Landgebruik, economische prestaties en de milieu- en sociaal-economische impacts van bio-energie zijn direct gerelateerd aan de dynamiek van de gehele agrarische sector (inclusief het gebruik van grasland en de veehouderij). Daarom dienen de effecten van de introductie van bio-energiegewassen en de ontwikkelingen van andere landgebruiksfuncties integraal geanalyseerd te worden. In dit proefschrift wordt een eerste stap gezet door de gehele agrarische sector (inclusief bio-energie gewassen) mee te nemen in de spatiotemporele analyse van broeikasgasemissies. Deze benadering kan ook worden toegepast voor de analyse andere impacts. Voor een betere betrouwbaarheid van een ex ante evaluatie van de 'go' en 'no go' gebieden voor bio-energieproductie, is een betere kwaliteit van (ruimtelijke-) gegevens vereist in termen van resolutie, nauwkeurigheid en documentatie. Met name data van huidig landgebruik (inclusief weilanden hun productiviteit en de intensiteit van het gebruik), agro-ecologische geschiktheid, bodemeigenschappen en klimaat zijn noodzakelijk. Bioraffinage van biomassa voor diverse toepassingen, zoals voedsel, veevoeder, vezels, energie en chemicaliën biedt mogelijkheden voor efficiënt gebruik van biomassa. Mogelijkheden van innovatieve routes moeten worden onderzocht om het verbeterpotentieel in het totale rendement en de economische prestaties van biomassa ketens te kwantificeren.

338

Samenvatting en conclusies

7.6

Markt en beleidsaanbevelingen •









Om een grote inzet van biomassa voor energie te realiseren en tegelijkertijd indirecte veranderingen in landgebruik (ILUC) te voorkomen, zal een verhoogde productie van bio-energie, diervoeder en voedsel moet worden gecompenseerd door verbeteringen in de landbouwproductiviteit. Daarnaast is de implementatie van duurzaamheids- en beleidskaders nodig om goed beheer van landgebruik en verbeteringen in de bosbouw, landbouw en veeteelt te waarborgen. Aanzienlijke bio-energie potentiëlen bevinden zich in minder ontwikkelde gebieden. Om de grote bio-energie potentiëlen in dergelijke regio's te ontwikkelen, is een meer geïntegreerde aanpak voor duurzame ontwikkeling nodig. Dit betekent niet alleen investeren in de agrarische sector, maar ook in onderwijs, infrastructuur en de ontwikkeling van markten. Aangezien ontwikkelingslanden vaak worden gekenmerkt door zwakke beleids- en institutionele kaders, brengt bio-energie productie in deze landen meer verantwoordelijkheden voor andere betrokken partijen (producenten, gebruikers, certificeringinstanties en overheden) met zich mee om duurzame bioenergieproductie te waarborgen. Een ex ante analyse van de beschikbaarheid van land, de economische haalbaarheid en de milieu-impacts staat identificatie van ‘go’ en ‘no-go’ gebieden voor bio-energie productie toe. Dit maakt op haar beurt een goede planning van landgebruik, duurzame investeringen in bio-energie productiecapaciteit en infrastructuur in de tijd mogelijk. Dergelijke inzichten kunnen ook investeerders en beleidsmakers helpen realistische schattingen van de economische haalbaarheid van een project te maken en het biedt de mogelijkheid om de randvoorwaarden om te voldoen aan duurzaamheidscriteria te definiëren. Dit kan helpen om investeringsrisico's te reduceren. Geavanceerde certificering van duurzame biomassa met verifieerbare en kwantificeerbare parameters vereist betere instrumentatie voor de beoordeling van milieu-impacts, economische prestaties en effecten op landgebruik. De gepresenteerde methoden in dit proefschrift kunnen wezenlijk bijdragen aan een ex ante analyse van de duurzaamheid van bio-energie productie voor certificering, maar betere data en fijnere modelresolutie zijn nodig. Idealiter wordt beleid inzake landbouw, milieu, duurzame energie en plattelandsontwikkeling op elkaar worden afgestemd om een duurzame bioenergie sector als onderdeel van duurzaam landgebruik en rurale ontwikkeling te realiseren.

339

References

References

341

SHADES OF GREEN

AAEA (2000). Commodity Costs and Returns Estimation Handbook: A Report of the AAEA Task Force on Commodity Costs and Returns. Iowa, USA, Agricultural & Applied Economics Association. Açucareira de Xinavane SA (2010). Small scale grower development project. A socio-economic develoment project in the Maputo province of Moçambique. . Xinavane, Açucareira de Xinavane SA. Agarwal, C., G. M. Green, J. M. Grove, T. P. Evans and C. M. Schweik (2001). A Review and Assessment of Land-Use Change Models: Dynamics of Space, Time, and Human Choice. washington D.C., United States Department of Agriculture, Indiana University Center for the Study of Institutions, Population, and Environmental Change. AgriHolland. (2009). "Dossier Landbouwgrond." Retrieved Januari 2009, 2009, from http://www.agriholland.nl/dossiers/landbouwgrond/home.html. Alcamo, J., G. J. J. Kreileman, M. S. Krol and G. Zuidema (1994). "Modeling the global societybiosphere-climate system: Part 1: Model description and testing." Water, Air, & Soil Pollution 76(1): 1-35. Alcamo, J., R. Leemans and G. J. J. Kreileman (1998). Global change scenarios of the 21st century. Results from the IMAGE 2.1 model. London, Elsevier Science. Alkemade, R., M. van Oorschot, L. Miles, C. Nellemann, M. Bakkenes and B. ten Brink (2009). "GLOBIO3: A Framework to Investigate Options for Reducing Global Terrestrial Biodiversity Loss." Ecosystems 12(3): 374-390. Allan, R. G., L. S. Pereira, D. Raes and M. Smith (1998). Crop evapotranspiration - Guidelines for computing crop water requirements - FAO Irrigation and drainage paper 56. Rome, Italy, FAO. ANE (2010). Roads of Mozambique. Maputo, Mozambique, Administração Nacional de Estradas. Animal science Group (2005). Kwantitatieve Informatie Veehouderij 2005-2006. A. s. group. Lelystad, Anmal Science Group: 414. Animal Science Group (2008). Handboek Snijmais. Lelystad, Animal Science Group, Prakrijkonderzoek plant en omgeving, Waganingen University and research centre. Batidzirai, B., A. P. C. Faaij and E. M. W. Smeets (2006). "Biomass and bioenergy supply from Mozambique." Energy for sustainable development X(1): 28. Batjes, N. H. (2005). SOTER-based soil parameter estimates for Central and Eastern Europe. Wageningen, The Netherlands, ISRIC, World Soil Information. Baumann, M., T. Kuemmerle, M. Elbakidze, M. Ozdogan, V. C. Radeloff, N. S. Keuler, A. V. Prishchepov, I. Kruhlov and P. Hostert (2011). "Patterns and drivers of post-socialist farmland abandonment in Western Ukraine." Land Use Policy 28(3): 552-562. Bellamy, P. E., P. J. Croxton, M. S. Heard, S. A. Hinsley, L. Hulmes, S. Hulmes, P. Nuttall, R. F. Pywell and P. Rothery (2009). "The impact of growing miscanthus for biomass on farmland bird populations." Biomass and Bioenergy 33(2): 191-199. Benjamin, K., G. Domon and A. Bouchard (2005). "Vegetation Composition and Succession of Abandoned Farmland: Effects of Ecological, Historical and Spatial Factors." Landscape Ecology 20(6): 627-647. Berghout, N. A. (2008). Technological learning in the German biodiesel industry: an experience curve approach to quantify reductions in production costs, energy use and greenhouse gas emissions. Science, Technology & Society, Copernicus Institute. Utrecht, Utrecht University. MSc. Bergsma, G., J. Vroonhof and V. Dornburg (2007). A greenhouse gas calculation methodology for biomass-based electricity, heat and fuels- the view of the Cramer Commission. Delft, CE Delft, Utrecht University: 39. Beringer, T., W. Lucht and S. Schaphoff (2011). "Bioenergy production potential of global biomass plantations under environmental and agricultural constraints." Global Change Biology Bioenergy 3(4): 299-312. 342

References

Berkhout, P. and S. van Berkum (2005). Hervorming suikermarktordening Gevolgen voor de suikerproducenten, de suikerverwerkers en de (sucro)chemische industrie. Den Haag, LEI. Berndes, G. (2002). "Bioenergy and water--the implications of large-scale bioenergy production for water use and supply." Global Environmental Change 12(4): 253-271. Berndes, G., M. Hoogwijk and R. van den Broek (2003). "The contribution of biomass in the future global energy supply: a review of 17 studies." Biomass and Bioenergy 25(1): 1-28. Bessembinder, J. J. E., P. A. Leffelaar, A. S. Dhindwal and T. C. Ponsioen (2005). "Which crop and which drop, and the scope for improvement of water productivity." Agricultural Water Management 73(2): 113-130. Bias, C. and C. Donovan (2003). Gaps and opportunities for agricultural sector development in Mozambique Maputo, Mozambique, National Agricultural Research Institute (INIA), Ministry of Agriculture and Rural Development; International Development, Department of Agricultural Economics,Michigan State University, East Lansing, Michigan: 129. Biewinga, E. E. and G. v. d. Bijl (1996). Sustainability of energy crops in Europe, a methodology developed and applied. Utrecht, CLM, Centre for Agriculture & Environment: 124 pages + appendices. Blottnitz, H. v. and M. A. Curran (2007). "A review of assessments conducted on bio-ethanol as a transportation fuel from a net energy, greenhouse gas, and environmental life cycle perspective." Journal of Cleaner Production 15(7): 607-619. BNDES and CGEE (2008). Sugarcane-based bioethanol: energy for sustainable development. Rio de Janeiro, BNDES. Boehmel, C., I. Lewandowski and W. Claupein (2008). "Comparing annual and perennial energy cropping systems with different management intensities." Agricultural Systems 96(1-3): 224236. Bogovin, A. V. (2001, April 2009). "Country Pasture/Forage Resource Profiles Ukraine." Retrieved April, 2009, from http://www.fao.org/ag/AGP/AGPC/doc/Counprof/Ukraine/ukraine.htm. Börjesson, P. (1999a). "Environmental effects of energy crop cultivation in Sweden--I: Identification and quantification." Biomass and Bioenergy 16(2): 137-154. Börjesson, P. (1999b). "Environmental effects of energy crop cultivation in Sweden--II: Economic valuation." Biomass and Bioenergy 16(2): 155-170. Boston Consultancy Group (1972). Perspectives on experience. Bouwman, A. F., K. W. Van der Hoek, B. Eickhout and I. Soenario (2005). "Exploring changes in world ruminant production systems." Agricultural Systems 84(2): 121-153. Bradley, D., F. Diesenreiter, M. Wild and E. Tromborg (2009). World Biofuel Maritime Shipping Study. Ottawa, Vienna, Aas, IEA Task 40. Brehmer, B., R. M. Boom and J. P. M. Sanders (2009). "Maximum fossil fuel feedstock replacement potential of petrochemicals via biorefineries." Chemical Engineering Research and Design In Press. Breux, J. and M. E. Salassi (2005). Projected Commodity Costs And Returns - Sugarcane Production in Louisiana, Farm Management Research & Extension Department of Agricultural Economics & Agribusiness. Britz, W. and A. Leip (2009). "Development of marginal emission factors for N losses from agricultural soils with the DNDC-CAPRI meta-model." Agriculture, Ecosystems & Environment 133(3-4): 267-279. Britz, W., P. H. Verburg and A. Leip (2011). "Modelling of land cover and agricultural change in Europe: Combining the CLUE and CAPRI-Spat approaches." Agriculture Ecosystems & Environment 142(1-2): 40-50. Brouwer, C. and M. Heibloem (1986). Irrigation Water Management: Irrigation Water Needs. Training manual no. 3. Part I - Principles of irrigation water needs. Rome, Italy, FAO, Natural Resources management and Environement department. 343

SHADES OF GREEN

Brouwer, F., H. L. Boogaard and R. C. M. Merkelbach (2003). Waterkansenkaart voor de functionele landbouw. een methode voor het vaststellen van geschiktheden en kwetsbaarheden van relevante gewasteelten voor Noord Nederland. Wageningen, Alterra, research instituut voor de Groene Ruimte: 91. Brouwer, F. and J. T. M. Huinink (2002). Opbrengstdervingspercentages voor combinaties van bodemtypen en grondwatertrappen. Geactualiseerde help-tabellen en opbrangstdepressiekaarten. Wageningen, Alterra, Research instituut groene ruimte en expertise centrum LNV: 38. Bruce, N., E. Rehfuess, S. Mehta, G. Hutton and K. Smith (2006). Indoor air pollution. Disease Control Priorities in Developing Countries, 2nd edition. D. T. Jamison, J. G. Breman, A. R. Meashamet al. Washington DC, The International Bank for Reconstruction and Development/The World Bank Group. Buishand, T. A. and C. A. Velds (1980). Neerslag en Verdamping. de Bilt, Koninklijk Nederlands Meteorologisch Instituut. Bullard, M. (2001). Economics of Miscanthus Production. Miscanthus for energy and fibre. M. B. Jones and M. Walsh. London, James & James (Science publishers) Ltd: 192. Bullard, M. and P. Matcalfe (2001). Estimating the energy requirements and CO2 emissions from production of the perrenial grasses miscanthus, switchgrass and reed canary grass., ADAS consulting Ltd on behalf of the Department of trade and industry: 94. Buss, J. (2007). SESAME PRODUCTION IN NAMPULA BASELINE SURVEY REPORT. CLUSA, Nampula, Mozambique, Southern African Development Community. Chemonics and IFCD (2007). Fertilizer Supply and Costs in Africa. Washington, Chemonics International Inc International Center for Soil Fertility and Agricultural Development. Cherubini, F., N. D. Bird, A. Cowie, G. Jungmeier, B. Schlamadinger and S. Woess-Gallasch (2009). "Energy- and greenhouse gas-based LCA of biofuel and bioenergy systems: Key issues, ranges and recommendations." Resources, Conservation and Recycling 53(8): 434-447. Christian, D. G., E. Haase, H. Schwarz, C. Dalianis, J. C. Clifton-Brown and S. Cosentino (2001). Agronomy of Miscanthus. Miscanthus for energy ad Fibre. M. B. Jones and M. Walsh. London, James & James (Science Publishers) Ltd: 192. Christian, D. G. and A. B. Riche (1999). Establishing fuel specifications of non wood biomass crops. Rothemsted, IACR-Rothamsted EPR Ltd: 64. Clifton-Brown, J. C. and I. Lewandowski (2000). "Water Use Efficiency and Biomass Partitioning of Three Different Miscanthus Genotypes with Limited and Unlimited Water Supply." Ann Bot 86(1): 191-200. Clifton-Brown, J. C., P. F. Stampfl and M. B. Jones (2004). "Miscanthus biomass production for energy in Europe and its potential contribution to decreasing fossil fuel carbon emissions." Global Change Biology 10(4): 509-518. Codd, E. F. (1968). Cellular Automata. New York and London, Academic Press inc. Coeleman, B. T., B. Pijanowska, G. J. Kasper, J. K. Gigler, C. Sonneveld, W. Huisman, A. E., J. v. Doorn, A. Bos, R. v. d. Pluijm and M. d. Boer (1996). Mogelijkheden voor kleinschalige energieopwekking met geteelde bioamssa. Wageingen, CPV, IMAG-DLO, ECN: 64. Commission of the European communities (2008). Directive of the European parliament and of the council on the promotion of the use of energy from renewable sources: 40 +appendices. Conselho de Ministros da república de Moçambique (2009). Publicação oficial da república de Moçambique, Política e Estratégia de Biocobustíveis. Resolução˚22/2009. n. Maputo, Moçambique. Costa Due. (2009). "Costa Due: Concrete stappen naar een duurzame Eemsmond." 2009, from http://www.costa-due.nl/.

344

References

Coughlin, P. E. (2006). Agricultural intensification Mozambique, infrastructure policy and institutopnal framework, when do problems signal opportunities? Maputo, EconPolicy research group lda: 78. Council of the European Communities (1991). Council directive of 12 December 1991 concerning the protection of waters against pollution caused by nitrates from agricultural sources. 91/676/EEC CPI (2009). Cost factors oin Mozambique. Maputo, Mozambique, Investment Promotion Centre: 16. Crago, C. L., M. Khanna, J. Barton, E. Giuliani and W. Amaral (2010). "Competitiveness of Brazilian sugarcane ethanol compared to US corn ethanol." Energy Policy 38(11): 7404-7415. Cramer, J. (2007). Testing framework for sustainable biomass; Final report from the project group “Sustainable production of biomass”, Commissioned by the Dutch Energy Transition’s Interdepartmental Programme Management (IPM): 60. Cuvilas, C. A., R. Jirjis and C. Lucas (2010). "Energy situation in Mozambique: A review." Renewable and Sustainable Energy Reviews 14(7): 2139-2146. Danalatos, N. G., S. V. Archontoulis and I. Mitsios (2007). "Potential growth and biomass productivity of Miscanthus×giganteus as affected by plant density and N-fertilization in central Greece." Biomass and Bioenergy 31(2-3): 145-152. Dastane, N. G. (1978). Effective rainfall in irrigated agriculture Rome, Italy: FAO. de Fraiture, C. and G. Berndes (2009). Biofuels and Water. Biofuels: Environmental Consequences and Interactions with Changing Land Use. Proceedings of the Scientific Committee on Problems of the Environment (SCOPE) International Biofuels Project Rapid Assessment, 2225 September 2008, Gummersbach Germany. R. W. Howarth and S. Bringezu. Ithaca NY, USA., Cornell University: 139-153. de Jong, K. (2009). "Aguila documentation Release 1.1.0." Retrieved January 2011, 2011, from http://pcraster.sourceforge.net/Aguila/1.1.0/. de Vries, B. J. M., D. P. van Vuuren and M. M. Hoogwijk (2007). "Renewable energy sources: Their global potential for the first-half of the 21st century at a global level: An integrated approach." Energy Policy 35(4): 2590-2610. De Vries, S. C., G. W. J. Van De Ven, M. K. Van Ittersum and K. E. Giller (2011a). "The productionecological sustainability of cassava, sugarcane and sweet sorghum cultivation for bioethanol in Mozambique." GCB Bioenergy 4(1): 20-35. de Vries, S. C., G. W. J. van de Ven, M. K. van Ittersum and K. E. Giller (2011b). "Resource use efficiency and environmental performance of nine major biofuel crops, processed by firstgeneration conversion techniques." Biomass and Bioenergy In Press, Corrected Proof. de Wit, M. and A. Faaij (2010). "European biomass resource potential and costs." Biomass and Bioenergy 34(2): 188-202. de Wit, M., M. Junginger, S. Lensink, M. Londo and A. Faaij (2010). "Competition between biofuels: Modeling technological learning and cost reductions over time." Biomass and Bioenergy 34(2): 203-217. de Wit, M., M. Londo and A. Faaij (2011a). "Productivity developments in European agriculture: Relations to and opportunities for biomass production." Renewable and Sustainable Energy Reviews 15(5): 2397-2412. de Wit, M. P., M. Junginger and A. P. C. Faaij (2011b). "Learning in dedicated wood production systems: past trends, future outlook and implications for bioenergy." submitted. de Wit, M. P., D. Lessmann, M. Londo and A. P. C. Faaij (2011c). "Environmental impacts of integrating biomass production into European agriculture." Forthcomming. de Wolf, M. and A. van der Klooster (2006). Kwantitatieve Informatie Akkerbouw en Vollegrondsteelt 2006. Wageningen, PPO Wageningen Universiteit. DEFRA (2002). Growing short rotation coppice. Best practical guidelines for applicants to DEFRA'S energy crops scheme. London, DEFRA Department for Environment, Food & Rural affairs: 31. 345

SHADES OF GREEN

Del Gatto, F. (2003). Forest law enforcement in Mozambique: An Overview. Mission Report. Maputo, MINISTRY OF AGRICULTURENATIONAL DIRECTORATE OF FORESTRY AND WILDLIFE, FAO: 19 + Annexes. DIVA GIS. (2011). "Country specific free GIS data: Ukraine. Inland waters: Rivers, canals, and lakes. Seperate files for line and area features." Retrieved 5 January 2011, 2011, from http://www.diva-gis.org. DNTF (2008a). GIS map Areas Communidades Mozambique. Maputo, Mozambique, Direcção Nacional de Terras e Florestas. DNTF (2008b). GIS map Forest Concessions Mozambique. Maputo, Mozambique, Direcção Nacional de Terras e Florestas. DNTF (2008c). GIS map Parcelas Concedidas Mozambique. Maputo, Mozambique, Direcção Nacional de Terras e Florestas DNTF (2008d). GIS map Protected areas Mozambique. Maputo, Mozambique, Direcção Nacional de Terras e Florestas. DNTF (2008e). GIS map rivers Mozambique. Maputo, Mozambique, Direcção Nacional de Terras e Florestas DNTF, IIAM and CeneCarta (2008). Inventário de terra disponível para investmento agrário no país. Maputo, DNTF. IIAM, CeneCarta. Dornburg, V. and A. P. C. Faaij (2001). "Efficiency and economy of wood-fired biomass energy systems in relation to scale regarding heat and power generation using combustion and gasification technologies." Biomass and Bioenergy 21(2): 91-108. Dornburg, V., A. P. C. Faaij, P. A. Verweij, M. Banse, K. van Diepen, H. van Keulen, H. Langeveld, M. Meeusen, G. van de Ven, F. Wester, R. Alkemade, B. ten Brink, G. J. van den Born, M. van Oorschot, J. Ros, F. Smout, D. van Vuuren, R. van den Wijngaart, H. Aiking, M. Londo, H. Mozaffarian and K. Smekens (2008). Scientific Assessment and Policy Analysis for Climate Change (WAB) Biomass Assessment Assessment of global biomass potentials and their links to food, water, biodiversity, energy demand and economy. E. Lysen and S. v. Egmond. Bilthoven, MNP, Universiteit Utrecht, Wageningen Universiteit en research centre, Vrije universiteit Amsterdam, ECN, UCE: 108. Dornburg, V., D. van Vuuren, G. van de Ven, H. Langeveld, M. Meeusen, M. Banse, M. van Oorschot, J. Ros, G. J. van den Born, H. Aiking, M. Londo, H. Mozaffarian, P. Verweij, E. Lysen and A. Faaij (2010). "Bioenergy revisited: Key factors in global potentials of bioenergy." Energy & Environmental Science 3(3): 258-267. Duinkerken, G. v. (2007). De voederconversie van melkvee (Rapport / Animal Sciences Group ) (Lelystad - Animal Sciences Group ). EC (2008). Directive of the European parliament and of the council on the promotion of the use of energy from renewable sources: 40 +appendices. EC (2009a). Directive 2009/28/EC of the European Parliament and of the council on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC. EC (2009b). Guidance document. The Application of the High Nature Value Impact Indicator 20012013, European Commission, Agriculture and rural development. The European Evaluation Network for Rural Development. Econergy (2008). Mozambique Biofuels Assessment - Final Report. Washington, DC, Econergy International Corporation, Prepared for the Ministry of Agriculture of Mozambique and the Ministry of Energy of Mozambique. : 426 + annexes. Edwards, R., J. F. Larivé, V. Mahieu and P. Rouveirolles (2007). Well-to Wheel analysis of future automotive fuels and powertrains in the European context. Well-to-Tank report Version 2c, March 2007, JRC, EU: 140.

346

References

EEA (2006). How much bioenergy can Europe produce without harming the environment? Copenhagen, Denmark, EEA: 67. EEA (2007). Estimating the environmentally compatible bio-energy potential from agriculture. Copenhagen, Denmark, EEA: 134. Eggers, J., K. Tröltzsch, A. Falcussi, L. Maiorano, P. H. Verburg, E. Framstad, G. Louette, D. Maes, S. Nagy, W. Ozinga and B. Belbaere (2009). "Is biofuel policy harming biodiversity in Europe?" Global Change Biology Bioenergy 1: 18-34. Eickhout, B. and A. G. Prins (2008). Eururalis 2.0 Technical Background and indicator documentation. Wageingen, Bilthoven, Wageningen UR, MNP: 84. Elbersen, B. S. and M. van Eupen (2008). Landbouwgrond met hoge natuurwaarden in Nederland op de kaart. Wageningen, Alterra: 125. Elbersen, W. H. (2008). Personal communication. AFSG, WUR, Wageningen. Elbersen, W. H., R. R. Bakker and B. S. Elbersen (2005). A simple method to estimate practical field yields of biomass grasses in Europe. 14th European Biomass Conference Paris, France. Elsayed, M. A., R. Matthews and N. D. Martimer (2003). Carbon and energy balances for a range of biofuels options Sheffield, UK, Sheffield Hallam University, Recources research unit: 25 + Appendices. Energy Valley. (2009). "Energy Valley: Energy for Life, Energy for Business." 2009, from http://www.energyvalley.nl/. Eppink, L. A. A. J. and W. P. Spaan (1989). "Agricultural wind erosion measures in the Netherlands." Soil technology Series 1: 1-13. Ercoli, L., M. Mariotti, A. Masoni and E. Bonari (1999). "Effect of irrigation and nitrogen fertilization on biomass yield and efficiency of energy use in crop production of Miscanthus." Field Crops Research 63(1): 3-11. Ericsson, K. and L. J. Nilsson (2006). "Assessment of the potential biomass supply in Europe using a resource-focused approach." Biomass and Bioenergy 30(1): 1-15. ESA (2005). IONIA GLOBCOVER Project, MEDIAS-France, ESA. ESA GlobCover Project, l. b. (2005). IONIA GLOBCOVER, MEDIAS-France, ESA. ESR and USDA (2011). Fertilizer Use and Price Economic research services, USDA, Washington. European Commission (2001). COUNCIL REGULATION (EC) No 2501/2001 of 10 December 2001 applying a scheme of generalised tariff preferences, European Commission, Brussels. European Commission (2006). Biofuels in the European Union. A vision for 2030 and beyond. Brussels, Final report of the Biofuels Research Advisory Council, European Commission,. European Union (2007). Biofuels: aid per hectare of energy crops reduced as the area exceeds 2 million hectares Press release IP/07/1528 17 October 2007. Eurostat. (2008). from http://epp.eurostat.ec.europa.eu/. Evers, A. (2008). Personal communication. Animal Science group, WUR, Wageningen,. Faaij, A., I. Steetskamp, A. van Wijk and W. Turkenburg (1998). "Exploration of the land potential for the production of biomass for energy in the Netherlands." Biomass and Bioenergy 14(5-6): 439-456. Faaij, A. P. C. and H. M. Junginger (2010). Chapter 10, Bioenergy. Technological Learning in the energy sector, lessons for policy, industry and science. H. M. Junginger, W. van Sark and A. P. C. Faaij. Northampton, Massachusetts, USA, Edward Elgar. FAO - GIS UNIT (2007). Combined suitability of global land area for pasture and rainfed crops (low input level) (FGGD). Rome, Italy. FAO (2000). Global map of population density. FAO (2003a). Forestry outlook study for Africa : regional report - opportunities and challenges towards 2020. FAO forestry paper. Rome, Food and Agricultural Organization of the United Nations

347

SHADES OF GREEN

FAO (2003b). World agriculture towards 2015/2030 an FAO perspective. J. Bruinsma. Rome, Food and Agriculture Organisation. FAO (2005a). FGGD 2.3 - Global population density estimates 2015. Rome, Food Insecurity, Poverty and Environment Global GIS Database (FGGD). FAO (2005b). Global forest resource assessment, Progress towards sustainable forest management. Rome, Food and Agriculture Organization of the United Nations FAO (2005c). Irrigation in Africa in figures. K. Frenken. Rome, Italy, Food and Agricultural organisation of the United Nations, Land and Water Development Division: 633. FAO (2005d). Livestock sector brief Mozambique. Rome, Livestock information, sector analysis and policy branch. Food and Agricultural Organisation.: 15. FAO (2005e). Predicted global cattle density (2005), corrected for unsuitability, adjusted to match FAOSTAT totals for the year 2005. Rome, FAO, Animal Production and Health Division The methodology and sources of data are fully described in a document entitled: "The Gridded Livestock of the World FAO (2007)". . FAO (2009a). CROPWAT 8.0. Rome, Italy, Land and Water Development Division of FAO: software tool to calculate evapotranpiration and irrigation requirements of crops. FAO. (2009b). "Food price Index." Retrieved January 2009, 2009, from http://www.fao.org/worldfoodsituation/FoodPricesIndex/en/. FAO (2010a). FAOSTATstatistical database, Food and Agricultural Organization. Rome. FAO (2010b). GLOBAL FOREST RESOURCES ASSESSMENT 2010 COUNTRY REPORT Mozambique. Rome, Forestry Department, Food and Agriculture Organization of the United Nations: 54. FAO (2010c). GLOBAL FOREST RESOURCES ASSESSMENT 2010 COUNTRY REPORT Ukraine, Ukraine, State statistic service. FAO (2011). Agriculture towards 2050, Outlook for Other Eastern European Countries, United Nations Food and Agriculture Organisation. FAO and IIASA (2000a). AEZ Global suitability of irrigated sugar crops (high inputs), FAO IIASA. FAO and IIASA (2000b). Agro-Ecological Zone Methodology. FAO and IIASA (2007a). Suitability of global land area for pasture (FGGD). Rome, FAO: The FGGD pasture suitability map is a global raster datalayer with a resolution of 5 arc-minutes. Each pixel contains the suitability index for pasture, 2005 version. FAO and IIASA (2007b). Suitability of global land area for rainfed crops, using maximising crop and technology mix (FGGD). FAO and WFP (2010). Crop and food security assessment mission to Mozambique, special report. Rome, Food and Agriculture Organisation of the United Nations. World Food Programme. : 36. Fehrenbach, H., J. Giegrich, G. A. Reinhardt, J. Schmitz, U. Sayer, M. Gretz, E. Seizinger and K. Lanje (2008). Criteria for a sustainable use of bioenergy on a global scale. Dessau-Rosslau, Istitut fur energie- und umweltforschung Heidelberg, FSC Arbeitsgruppe Deutschland Freiburg, German Watch; on behalf of the Federal Environmental Agency: 128 + appendices. Fischer, G., E. Hizsnyik, S. Prieler and H. T. Velthuizen (2007). Assessment of biomass potentials for biofuel feedstock production in Europe: Methodology and results. Laxenburg, Austria, IIASA: 75. Fischer, G., S. Prieler, H. van Velthuizen, G. Berndes, A. Faaij, M. Londo and M. de Wit (2010a). "Biofuel production potentials in Europe: Sustainable use of cultivated land and pastures, Part II: Land use scenarios." Biomass and Bioenergy 34(2): 173-187. Fischer, G., S. Prieler, H. van Velthuizen, S. M. Lensink, M. Londo and M. de Wit (2010b). "Biofuel production potentials in Europe: Sustainable use of cultivated land and pastures. Part I: Land productivity potentials." Biomass and Bioenergy 34(2): 159-172. Fischer, G., H. T. van Veldhuizen and F. Nachtergaele (2000). Global Agro-Ecological Zones Assessment: Methodology and Results. Laxenburg, Austria, IIASA: 338. 348

References

Foereid, B., A. de Neergaard and H. Høgh-Jensen (2004). "Turnover of organic matter in a Miscanthus field: effect of time in Miscanthus cultivation and inorganic nitrogen supply." Soil Biology and Biochemistry 36(7): 1075-1085. Foley, J. A., R. de Fries, G. P. Asner, C. Barford, G. Bonan, S. R. Carpenter, F. S. Chapin, M. T. Coe, G. C. Daily, H. K. Gibbs, J. H. Helkowski, T. Holloway, E. A. Howard, C. J. Kucharik, C. Monfreda, J. A. Patz, I. C. Prentice, N. Ramankutty and P. K. Snyder (2005). "Global consequences of land use." Science 309(5734): 570-574. Fox, L., E. Bardasi and K. V. d. Broeck (2005). Poverty in Mozambique: Unraveling Changes and Determinants. Africa Region Working Paper Series 87. Wahington, World Bank: 55. Fraters, B., L. J. M. Boumans, T. C. van Leeuwen and J. W. Reijs (2007). De uitspoeling van het stikstofoverschot naar grond- en oppervlaktewater op landbouwbedrijven. Bilthoven, RIVM, LEI-Wageningen UR: 62 + Appendices. Gallagher, E. (2008). The Gallagher Review of the indirect effects of biofuels production. East Sussex, Renewable Fuels Agency: 90. GBEP (2011). The global bioenergy partnership sustainability indicators for bioenergy. First edition. Rome, GBEP secretariat FAO GDS (2005). Value chain analysis for strategic sectors in Mozambique. Reston, Virginia, Global Development Solution, LLC. Gerbens-Leenes, W., A. Y. Hoekstra and T. H. van der Meer (2009). "The water footprint of bioenergy." Proceedings of the National Academy of Sciences 106(25): 10219–10223. GF Energy (2010). Personal communication wood pellet trade East Africa - Rotterdam Harbour Groom, M. J., E. M. Gray and P. A. Townsend (2008). "Biofuels and biodiversity: Principles for creating better policies for biofuel production." Conservation Biology 22(3): 602-609. Haberl, H., T. Beringer, S. C. Bhattacharya, K. H. Erb and M. Hoogwijk (2010). "The global technical potential of bio-energy in 2050 considering sustainability constraints." Current Opinion in Environmental Sustainability 2(5-6): 394-403. Hamelinck, C. N. (2004). Outlook for advanced biofuels. Copernicus Institute, Beta faculty. Utrecht, Utrecht University: 232. Hamelinck, C. N. and A. P. C. Faaij (2006). "Outlook for advanced biofuels." Energy Policy 34(17): 3268-3283. Hamelinck, C. N. and M. Hoogwijk (2007). Future scenarios for first and second generation biofuels. Utrecht Bilthoven, Ecofys MNP: 78. Hamelinck, C. N., G. v. Hooijdonk and A. P. C. Faaij (2005a). "Ethanol from lignocellulosic biomass: techno-economic performance in short-, middle- and long-term." Biomass and Bioenergy 28(4): 384-410. Hamelinck, C. N., R. A. A. Suurs and A. P. C. Faaij (2003). International bioenergy transport costs and energy balance. Utrecht, The Netherlands, Copernicus Institute, department of Energy, Science and Technology, Utrecht University. Hamelinck, C. N., R. A. A. Suurs and A. P. C. Faaij (2005b). "International bioenergy transport costs and energy balance." Biomass and Bioenergy 29(2): 114-134. Hamelinck, C. N. and R. van den Broek (2005). Participative LCA on biofuels Utrecht, SenterNovem: 128. Heijboer, D. and J. Nellestein (2002). Klimaatatlas van Nederland. De normaalperiode 1971-2000. de Bilt, Elmar multimedia. Heller, M. C., G. A. Keoleian, M. K. Mann and T. A. Volk (2004). "Life cycle energy and environmental benefits of generating electricity from willow biomass." Renewable Energy 29(7): 1023-1042. Heller, M. C., G. A. Keoleian and T. A. Volk (2003). "Life cycle assessment of a willow bioenergy cropping system." Biomass & Bioenergy 25(2): 147-165. Hellmann, F. and P. H. Verburg (2008). "Spatially explicit modelling of biofuel crops in Europe." Biomass and Bioenergy. 349

SHADES OF GREEN

Helsel, Z. R., Ed. (1997). Energy in Plant Nutrition and Pest control. Amsterdam, Elsevier. Hettinga, W. G., H. M. Junginger, S. C. Dekker, M. Hoogwijk, A. J. McAloon and K. B. Hicks (2009). "Understanding the reductions in US corn ethanol production costs: An experience curve approach." Energy Policy 37(1): 190-203. Himken, M., J. Lammel, D. Neukirchen, U. Czypionka-Krouse and H. W. Wolfs (1997). "Cultivation of Miscanthus under West European conditions: Seasonal changes in dry matter production, nutrient uptake and remobilization." Plant and soil 189: 117-126. Hoefnagels, R., E. Smeets and A. Faaij (2010). "Greenhouse gas footprints of different biofuel production systems." Renewable and Sustainable Energy Reviews 14(7): 1661-1694. Hoogeveen, M. W., K. H. M. v. Bommel and G. Cotteleer (2003). Beregening in land- en tuinbouw. Rapport voor de Droogtestudie Nederland. Den Haag, LEI: 64. Hooghart, J. C. and W. N. Lablans (1988). Van Penman naar Makkink: een nieuwe berekeningswijze voor de klimatologische verdampingsgetallen. De Bilt, KNMI: 67. Hoogwijk, M., A. Faaij, B. de Vries and W. Turkenburg (2009). "Exploration of regional and global cost-supply curves of biomass energy from short-rotation crops at abandoned cropland and rest land under four IPCC SRES land-use scenarios." Biomass and Bioenergy 33(1): 26-43. Hoogwijk, M., A. Faaij, B. Eickhout, B. de Vries and W. Turkenburg (2005). "Potential of biomass energy out to 2100, for four IPCC SRES land-use scenarios." Biomass and Bioenergy 29(4): 225-257. Horner, M. W., T. Zhao and T. S. Chapin (2011). "Toward an Integrated GIScience and Energy Research Agenda." Annals of the Association of American Geographers 101(4): 764-774. Houghton, R. A. (1999). "The annual net flux of carbon to the atmosphere from changes in land use 1850–1990*." Tellus 51B: 298–313. Houtsma, J. (2008). Personal Communication. Da Vinci Finance, Friesland. Huisman, W., P. Venturi and J. Molenaar (1997). "Costs of supply chains of Miscanthus giganteus." Industrial Crops and Products 6(3-4): 353-366. IEA (2010a). Energy balances of non-OECD countries. Paris, Internation Energy Agency. IEA (2010b). World Energy Outlook, International Energy Agency: 736 IEA and OECD (2011). World Energy Outlook 2011. Paris, France, International Energy Agency IMF (2007). Republic of Mozambique: Selected Issues. Washington DC, IMF. IMF (2010). World Economic Outlook Database, International Monetary Fund, Washington, D.C. INE (2002). Censo Agro-pecuário 1999-2000, Apresentação sumária dos resultados Moçambique. Maputo, Instituto Nacional de Estatística: 105. INE (2003). Censo Agro-Pecuário 1999-2000, resultados Temáticos. Maputo, Instituto Nacional de Estatística. IPCC (2003). Good Practice Guidance for Land Use, Land-Use Change and Forestry. J. Penman, M. Gytarsky, T. Hiraishiet al. Hayama, Kanagawa, Japan, IPCC National Greenhouse Gas Inventories Programme Technical Support Unit, Institute for Global Environmental Strategies. IPCC (2006). IPCC Guidelines for National Greenhouse Gas Inventories, Prepared by the National Greenhouse Gas Inventories Programme, Eggleston H.S., Buendia L., Miwa K., Ngara T. and Tanabe K. (eds). IGES. Hayama, Kanagawa, Japan. IPCC (2007a). Climate Change 2007: Synthesis Report. Contribution of Working Groups I, II and III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Core Writing Team, R. K. Pachauri and A. Reisinger. Geneva, Switzerland, IPCC: 104. IPCC (2007b). Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the IPCC. Cambridge, Intergovernmental Panel on Climate Change. IPCC (2007c). Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the IPCC. Technical Summary. S. Solomon, D. Qin, M. 350

References

Manninget al. Cambridge, United Kingdom and New York, NY, USA., Cambridge University Press. IPCC (2011). Special Report on Renewable Energy Sources and Climate Change Mitigation. The Fifth Assessment Report. O. Edenhofer, R. Pichs-Madruga, Y. Sokona and et al. Cambridge, United Kingdom and New York, NY, USA, Cambridge University Press. ISCC. (2010). "Sustainability Requirements for the Production of Biomass." Retrieved December 2009, from http://www.iscc-project.org/project. ISO (2008). New work item proposal: International Standards for sustainability criteria for biofuel, ISO. IUCN and UNEP-WCMC (2010). World Database on Protected Areas UNEP World Conservation Monitoring Centre, Cambridge. Jansen, L. J. M., M. Bagnoli and M. Focacci (2008). "Analysis of land-cover/use change dynamics in Manica Province in Mozambique in a period of transition (1990-2004)." Forest Ecology and Management 254(2): 308-326. JEC (2008). Well-to-Wheels analysis of future automotive fuels and powertrains in the European context, Version 3., European Council for Automotive R&D (EUCAR), European association for environment, health and safety in oil refining and distribution (CONCAWE), the Institute for Environment and Sustainability of the EU Commission's Joint Research Centre (JRC/IES). Jelsma, I. (2010). Smallholder Sugarcane Production Systems in Xinavane, Mozambique: Report from the Field. Wageningen, Plant Production Systems, Plant Sciences Group, Wageningen University. JRC. (2009). "Input data relevant to calculating default GHG emissions from biofuels according to RE Directive Methodology." from http://re.jrc.ec.europa.eu/biof/html/input_data_ghg.htm. Junginger, M., E. de Visser, K. Hjort-Gregersen, J. Koornneef, R. Raven, A. Faaij and W. Turkenburg (2006). "Technological learning in bioenergy systems." Energy Policy 34(18): 4024-4041. Junginger, M., A. Faaij, R. Björheden and W. C. Turkenburg (2005). "Technological learning and cost reductions in wood fuel supply chains in Sweden." Biomass and Bioenergy 29(6): 399-418. Kahle, P., S. Beuch, B. Boelcke, P. Leinweber and H.-R. Schulten (2001). "Cropping of Miscanthus in Central Europe: biomass production and influence on nutrients and soil organic matter." European Journal of Agronomy 15(3): 171-184. Kaltschmitt, M. and G. A. Reinhardt (1997). Nachwachsende Energietrager. Grundlagen, verfahren, okologische Bilanzierung. Branschweig Wiesbaden, Friedr. Vieweg & Sohn Verlagsgesellschaft. Kaltschmitt, M., G. A. Reinhardt and T. Stelzer (1997). "Life cycle analysis of biofuels under different environmental aspects." Biomass and Bioenergy 12(2): 121-134. Karekezi, S. (2002). "Renewables in Africa - meeting the energy needs of the poor." Energy Policy 30(11-12): 1059-1069. Karssenberg, D., O. Schmitz, P. Salamon, K. de Jong and M. F. P. Bierkens (2010). "A software framework for construction of process-based stochastic spatio-temporal models and data assimilation." Environmental Modelling & Software 25(4): 489-502. Kempenaar, C., L. v. d. Brink, C. B. Bus, J. A. M. Groten, C. L. M. d. Visser and L. A. P. Lotz (2003). Gangbare landbouwkundige praktijk en recente ontwikkelingen voor vier akkerbouwgewassen in Nederland. Wageingen, Plant Reasearch International, BU Gewas & Productie-ecologie, Wageningen. Praktijkonderzoek Plant & Omgeving, Cluser akkerbouw, groene ruimte & vollegrondsgroente, Lelystad: 38. Khanna, M., B. Dhungana and J. Clifton-Brown (2008). "Costs of producing miscanthus and switchgrass for bioenergy in Illinois." Biomass and Bioenergy 32(6): 482-493. Kim, H., S. Kim and B. E. Dale (2009). "Biofuels, Land Use Change, and Greenhouse Gas Emissions: Some Unexplored Variables." Environmental Science & Technology 43(3): 961-967. KNMI (2002). Klimatatlas; Langjarige gemiddelden en extremen, tijdvak 1971-2000 KNMI. 351

SHADES OF GREEN

Kort, J., M. Collins and D. Ditsch (1998). "A review of soil erosion potential associated with biomass crops." Biomass and Bioenergy 14(4): 351-359. Krasuska, E., C. Cadórniga, J. L. Tenorio, G. Testa and D. Scordia (2010). "Potential land availability for energy crops production in Europe." Biofuels, Bioproducts and Biorefining 4(6): 658-673. Kucher, O. (2007). Ukrainian Agriculture and Agri-Environmental Concern. Wien, Universität für Bodenkultur Wien Department für Wirtschafts- und Sozialwissenschaften. Kuemmerle, T., O. Chaskovskyy, J. Knorn, V. C. Radeloff, I. Kruhlov, W. S. Keeton and P. Hostert (2009). "Forest cover change and illegal logging in the Ukrainian Carpathians in the transition period from 1988 to 2007." Remote Sensing of Environment 113(6): 1194-1207. Kuemmerle, T., P. Olofsson, O. Chaskovskyy, M. Baumann, K. Ostapowicz, C. E. Woodcock, R. A. Houghton, P. Hostert, W. S. Keeton and V. C. Radeloff (2011). "Post-Soviet farmland abandonment, forest recovery, and carbon sequestration in western Ukraine." Global Change Biology 17(3): 1335-1349. Kuemmerle, T., V. C. Radeloff, K. Perzanowski and P. Hostert (2006). "Cross-border comparison of land cover and landscape pattern in Eastern Europe using a hybrid classification technique." Remote Sensing of Environment 103(4): 449-464. Kuikman, P. J., W. J. M. de Groot, R. F. H. Hendriks, J. Verhagen and F. de Vries (2003). Stocks of C in soils and emissions of CO2 from agricultural soils in the Netherlands. Wageningen, Alterra, PRI: 39 + appendices. Kuikman, P. J., J. J. H. van den Akker and F. de Vries (2005). Emissie van N2O en CO2 uit organische landbouwbodems. Wageningen, the Netherlands, Alterra: 66. Kuiper, L. (2003). Samenvatting van de resultaten van zes jaar onderzoek naar energieteelt. Wageningen, Cetrum voor Biomassa Innovatie. Laclau, J. P., P. Deleporte, J. Ranger, J. P. Bouilleti and G. Kazotti (2003). "Nutrient dynamics throughout the rotation of Eucalyptus clonal stands in Congo." Annals of Botany 91(7): 879892. Lambin, E. F., B. L. Turner, H. J. Geist, S. B. Agbola, A. Angelsen, J. W. Bruce, O. T. Coomes, R. Dirzo, G. Fischer, C. Folke, P. S. George, K. Homewood, J. Imbernon, R. Leemans, X. Li, E. F. Moran, M. Mortimore, P. S. Ramakrishnan, J. F. Richards, H. Skånes, W. Steffen, G. D. Stone, U. Svedin, T. A. Veldkamp, C. Vogel and J. Xu (2001). "The causes of land-use and land-cover change: moving beyond the myths." Global Environmental Change 11(4): 261-269. Lapola, D. M., R. Schaldach, J. Alcamo, A. Bondeau, J. Koch, C. Koelking and J. A. Priess (2010). "Indirect land-use changes can overcome carbon savings from biofuels in Brazil." Proceedings of the National Academy of Sciences 107(8): 3388-3393. LEI CBS (2007). Land- en tuinbouwcijfers 2007. Den Haag, LEI CBS: 270. Leip, A., G. Marchi, R. Koeble, M. Kempen, W. Britz and C. Li (2008). "Linking an economic model for European agriculture with a mechanistic model to estimate nitrogen and carbon losses from arable soils in Europe." Biogeosciences 5: 73–94. Lerman, Z., D. Sedik, N. Pugachev and A. Goncharuk (2006). Ukraine after 2000: A Fundamental Change in Land and Farm Policy? Rome, FAO. Lesschen, J. P. (2008). Policy innitiatives for climate change mitigation agricultural techniques. Deliverable 7: European quantification results. Wageningen, Alterra, Wageningen University and Research centre. Lesschen, J. P., G. L. Velthof, W. de Vries and H. Kros (2010). "Estimation of N2O emission factors for soils depending on N input sources and environmental conditions." Environmental Pollution. Lesschen, J. P., G. L. Velthof, W. de Vries and J. Kros (2011). "Differentiation of nitrous oxide emission factors for agricultural soils." Environmental Pollution 159(11): 3215-3222. Lewandowski, I., J. C. Clifton-Brown, J. M. O. Scurlock and W. Huisman (2000). "Miscanthus: European experience with a novel energy crop." Biomass and Bioenergy 19(4): 209-227.

352

References

Lewandowski, I. and A. P. C. Faaij (2006). "Steps towards the development of a certification system for sustainable bio-energy trade." Biomass and Bioenergy 30(2): 83-104. Lewandowski, I. and A. Heinz (2003). "Delayed harvest of miscanthus--influences on biomass quantity and quality and environmental impacts of energy production." European Journal of Agronomy 19(1): 45-63. Lewandowski, I. and U. Schmidt (2006). "Nitrogen, energy and land use efficiencies of miscanthus, reed canary grass and triticale as determined by the boundary line approach." Agriculture, Ecosystems & Environment 112(4): 335-346. Lewandowski, I., J. M. O. Scurlock, E. Lindvall and M. Christou (2003). "The development and current status of perennial rhizomatous grasses as energy crops in the US and Europe." Biomass and Bioenergy 25(4): 335-361. Long, S. P., C. V. Beale and P. K. Farage (2001). Resource Capture by Miscanthus. Miscanthus for energy and fibre. M. B. Jones and M. Walsh. London, James & James (science Publishers) Ltd: 192. Macedo, I. C. and J. E. A. Seabra (2008). Mitigation of GHG emissions using sugar cane bioethanol. Sugar cane ethanol, contributions to climate change mitigation and the environment. P. Zuurbier and J. v. d. Vooren. Wageningen, Wageningen Academic Publishers: 255. MacKenzie, C. (2006). Forest Governance in Zambézia, Mozambique: Chinese Takeawey! Report for Forum of NGOs of Zambezia (FONGZA), Mozambique: 88. Magaia, H. E., M. Freire, A. Monjana, O. Davolio Marani, A. Zazzerini, M. Durante and F. Cecconi (2005). "SELECTION OF NEW SUNFLOWER (Helianthus annuus L.) SYNTHETIC VARIETIES ADAPTED FOR PRODUCTION AREAS OF MOZAMBIQUE." HELIA 28(43): 69-76. Malca, J. and F. Freire (2006). "Renewability and life-cycle energy efficiency of bioethanol and bioethyl tertiary butyl ether (bioETBE): Assessing the implications of allocation." Energy 31(15): 3362-3380. Mangue, P. D. (2000). Review of the existing studies related to fuelwood and/or charcoal in Mozambique. Maputo, EC-FAO Partnership: 37. Maposse, I. c. C., J. P. Muir and A. A. Alage (2003). "Status of range and forage research in Mozambique." African Journal of Range & Forage Science 20(1): 63 - 68. Marques, P. V., D. Y. Sonoda, L. B. Zilio, C. E. O. Xavier, F. J. A. d. Arruda, J. H. M. Rosa and A. A. Vitt (2009). Custo de produção agrícola e industrial de açúcar e álocol no Brasil na Safra 2007/2008. São Paulo, Universidade de São Paulo - USP, Escola Superior de Agricultura “Luiz de Queiroz” - ESALQ, Departamento de Economia, Administração e Sociologia, Programa de Educação Continuada em Economia e Gestão de empresas - PECEGE. Mas, J. F., H. Puig, J. L. Palacio and A. Sosa-López (2004). "Modelling deforestation using GIS and artificial neural networks." Environmental Modelling & Software 19(5): 461-471. Meeuws, R. (2004). Mozambique - Trade and Transport facilitation audit. Rijswijk, The Netherlands, NEA, Transport research and training, commissioned by the World Bank: 91. Melillo, J. M., J. M. Reilly, D. W. Kicklighter, A. C. Gurgel, T. W. Cronin, S. Paltsev, B. S. Felzer, X. Wang, A. P. Sokolov and C. A. Schlosser (2009). "Indirect Emissions from Biofuels: How Important?" Science 326(5958): 1397-1399. Menkveld, M. (2007). Beoordeling werkprogramma Schoon en Zuinig; Effecten op energiebesparing, hernieuwbare energie en uitstoot van broeikasgassen. Petten, The Netherlands, ECN, MNP. Mertens, B. and E. Lambin (1999). "Modelling land cover dynamics: integration of fine-scale land cover data with landscape attributes." International Journal of Applied Earth Observation and Geoinformation 1(1): 48-52. Mertens, B. and E. F. Lambin (1997). "Spatial modelling of deforestation in southern Cameroon: Spatial disaggregation of diverse deforestation processes." Applied Geography 17(2): 143162.

353

SHADES OF GREEN

Ministerie van Economische zaken (2008). Energierapport 2008. Den Haag, The Netherlands, Ministerie van economische zaken, Directoraat-Generaal voor energie en telecom: 126. Ministerie van VROM (2007). Nieuwe energie voor het klimaat; werkprogramma schoon en zuinig. VROM. den Haag, The Netherlands, Ministerie van VROM, Ministerie van Economische Zaken; Ministerie van Verkeer en Waterstaat; Ministerie van Landbouw, Natuur en Voedselkwaliteit; Ministerie van Financiën; ministerie van Buitenlandse Zaken.: 60. Mitchell, D. (2008). A Note on Rising Food Prices. Washington, The World Bank. MNP (2006). Integrated modelling of global environmental change; An overview of IMAGE 2.4. A. F. Bouwman, T. Kram and K. Klein Goldewijk. Bilthoven, The Netherlands, MNP: 228. Monti, A., N. Di Virgilio and G. Venturi (2008). "Mineral composition and ash content of six major energy crops." Biomass and Bioenergy 32(3): 216-223. Morgan, R. P. C. (2005). Soil erosion and Conservation. Oxford, Blackwell Publishing Ltd. Morton, R., K. Sharp, B. Chomiak, N. Stepanets, O. Muliar and N. Oleshko (2005). FARM REFERENCE HANDBOOK FOR UKRAINE, USAID. Mosier, A., C. Kroeze, C. Nevison, O. Oenema, S. Seitzinger and O. van Cleemput (1998). "Closing the global N2O budget: nitrous oxide emissions through the agricultural nitrogen cycle." Nutrient Cycling in Agroecosystems 52(2): 225-248. Murphy, J. D. and K. McCarthy (2005). "Ethanol production from energy crops and wastes for use as a transport fuel in Ireland." Applied Energy 82(2): 148-166. Nakicenovic, N., O. Davidson, Gerald Davis, A. Grübler, T. Kram, E. L. L. Rovere, Bert Metz, T. Morita, W. Pepper, H. Pitcher, A. Sankovski, P. Shukla, R. Swart, R. Watson and Z. Dadi (2000). IPCC SPECIAL REPORT EMISSIONS SCENARIOS:A Special Report of IPCC Working Group III. Cambridge, U.K., , Intergovernmental Panel on Climate Change: 599. Namburete, H. E. S. (2006). Mozambique Bio-Fuels. African Green Revolution Conference. Oslo Norway. NASA and NGA (2000). Shuttle Radar Topography Mission (SRTM) 30m Digital Elevation Data, CGIAR Consortium for Spatial Information (CGIAR-CSI) NEA (2004). Factorkosten van het goederenvervoer: een analyse van de ontwikkeling in de tijd, 2e druk. Rijswijk, Transport onderzoek en -opleiding,Adviesdienst verkeer en vervoer: 59 + appendices. NEN (2009). NTA 8080, Sustainability criteria for biomass for energy purposes, NEN. Neukirchen, D., M. Himken, J. Lammel, U. Czypionka-Krause and H. W. Olfs (1999). "Spatial and temporal distribution of the root system and root nutrient content of an established Miscanthus crop." European Journal of Agronomy 11(3-4): 301-309. Nhancale, B., S. Mananze, N. Dista, I. Nhantumbo and D. Macqueen (2009). Small and medium forest enterprises in Mozambique. IIED Small and Medium Forest Enterprise Series London, UK., Centro Terra Viva and International Institute for Environment and Development. NOAA. (2008). "NOAA weather stations database. National Oceanic and Atmospheric Administration of the USA, Earth system research laboratory." from http://www.climate-charts.com/worldindex.html. NRCS (2006). WEQ management period method wind erosion model worksheet, NRCS. Obernberger, I. and G. Thek (2010). The pellet handbook. The production and thermal utilisation of biomass pellets. London, Washington DC, Earthscan. OECD and FAO (2011). Agricultural outlook 2011-2020, OECD Publishing and FAO. OECD and IEA (2008). World Energy Outlook 2008. Paris, France, IEA, OECD: 501 + annexes. Open Street Map (2010). Open street map, Europe, Ukraine, places. http://download.geofabrik.de/osm/europe/, Open Street Map,. Osborn, S. and M. A. Trueblood (2002). Agricultural productivity and efficiency in Rusia and Ukraine. Building on a decade of reform., Market and trades economic division,Econnomic Research Service, USDA. 354

References

Otte, M. J. and P. Chilonda (2002). Cattle and small ruminant production systems in sub-Saharan Africa A systematic review. Rome, Livestock Information Sector Analysis and Policy Branch, FAO Agriculture Department. Overmars, K. P., P. H. Verburg and T. Veldkamp (2007). "Comparison of a deductive and an inductive approach to specify land suitability in a spatially explicit land use model." Land Use Policy 24(3): 584-599. Paracchini, M. L., J.-E. Petersen, Y. Hoogeveen, C. Bamps, I. Burfield and C. van Swaay (2008). High Nature Value Farmland in Europe An estimate of the distribution patterns on the basis of land cover and biodiversity data. Luxembourg, European Comission Joint research Centre, European Environment Agency. PCRaster. (2010). "PCRaster internet site, Available online at: http://pcraster.geo.uu.nl." Retrieved December 2010. Pennsylvania State University Libraries (1997). Railroad ukraine. Pereira, C., R. Brouwer, M. Monjane and M. Falcao (2001). Charcoal Potential in Southern Africa CHAPOSA, Final report for Mozambique prepared by the Eduardo Mondelane University. Stockholm, Stockholm Environment Institute. Perlack, R. D., L. L. Wright, A. F. Turhollow and R. L. Graham (2005). BIOMASS AS FEEDSTOCK FOR A BIOENERGY AND BIOPRODUCTS INDUSTRY: THE TECHNICAL FEASIBILITY OF A BILLION-TON ANNUAL SUPPLY. Oak Ridge, Oak Ridge National Laboratory, U.S. Department of Agriculture. Pimentel, D., C. Harvey, P. Resosudarmo, K. Sinclair, D. Kurz, M. McNair, S. Crist, L. Shpritz, L. Fitton, R. Saffouri and R. Blair (1995). "Environmental and Economic Costs of Soil Erosion and Conservation Benefits." Science 267(5201): 1117-1123. Pimentel, D., D. Pimentel and T. W. Patzek (2008). Ethanol Production: Energy and Economic Issues Related to U.S. and Brazilian Sugarcane. Biofuels, Solar and Wind as Renewable Energy Systems, Springer Netherlands: 357-371. Platts. (2010). "Platts European Marketscan." Retrieved September 2011, from www.platts.com. Popp, A., H. Lotze-Campen, M. Leimbach, B. Knopf, T. Beringer, N. Bauer and B. Bodirsky (2011). "On sustainability of bioenergy production: Integrating co-emissions from agricultural intensification." Biomass and Bioenergy 35(2011): 4770-4780. Poppe, K. J. (2004). Het Bedrijven-Informatienet van A tot Z. Den Haag, The Netherlands, LEI: 94. PPO, M. d. Wolf and A. v. d. Klooster (2006). Kwantitatieve Informatie Akkerbouw en Vollegrondsteelt 2006. Wageingen, Wageningen Universiteit. Principle Energy. (2009). "http://www.prinenergy.com/output/page185.asp." Retrieved 1st June, 2011. Projectgroep 'Duurzame productie van Biomassa' (2006). Criteria voor duurzame biomassa productie, Task Force Energietranstie: 30. Pueyo, Y. and S. Beguería (2007). "Modelling the rate of secondary succession after farmland abandonment in a Mediterranean mountain area." Landscape and Urban Planning 83(4): 245-254. Reeves, D. W. (1997). "The role of soil organic matter in maintaining soil quality in continuous cropping systems." Soil and Tillage Research 43(1-2): 131-167. Reidsma, P., T. Tekelenburg, M. van den Berg and R. Alkemade (2006). "Impacts of land-use change on biodiversity: An assessment of agricultural biodiversity in the European Union." Agriculture, Ecosystems & Environment 114(1): 86-102. Renka, R. J. (1987). "Interpolatory tension splines with automatic selection of tension factors." Society for Industrial and Applied Mathematics 8(3): 393-415. Republic of Mozambique (2010). STRATEGIC PLAN FOR AGRICULTURAL DEVELOPMENT PEDSA 20102019, For an integrated, prosperous, competitive and sustainable agriculture sector Maputo, Republic of Mozambique, MINISTRY OF AGRICULTURE.

355

SHADES OF GREEN

República de Moçambique (2010). Relatório de vista aos projectos de biocombustiveis no país. . Maputo, Mozambique, Ministério da energia Drirecção Nacional de enrgias nova e renováveis. Reuters. (2009). "Principle Energy to start on ethanol plant in May." Retrieved August 1st, 2011, from http://uk.reuters.com/article/2009/04/01/principle-mozambiqueidUKWEB378320090401. Riksen, M. J. P. M. and J. de Graaff (2001). "On-site and off-site effects of wind erosion on European light soils." Land Degradation & Development 12(1): 1-11. Rodríguez, E., C. S. Morris, J. E. Belz, E. C. Chapin, J. M. Martin, W. Daffer and S. Hensley (2005). An Assessment of the SRTM Topographic Products, Technical Report JPL D-31639. Pasadena, California, JPL, NASA. Roelofs, P. (2008). Personal communication. Animal Science groeup, WUR, Wageningen. Romkes, P. F. A. M. and O. Oenema (2004). Quick scan soils in the Netherlands, overview of the soil status with reference to the forthcomming EU soil strategy. Wageningen, The Netherlands, Alterra: 96. Rosenqvist, H. and M. Dawson (2005). "Economics of willow growing in Northern Ireland." Biomass and Bioenergy 28(1): 7-14. Rounsevell, M. D. A., I. Reginster, M. B. Araújo, T. R. Carter, N. Dendoncker, F. Ewert, J. I. House, S. Kankaanpää, R. Leemans, M. J. Metzger, C. Schmit, P. Smith and G. Tuck (2006). "A coherent set of future land use change scenarios for Europe." Agriculture, Ecosystems & Environment 114(1): 57-68. Rowe, R. L., N. R. Street and G. Taylor (2009). "Identifying potential environmental impacts of largescale deployment of dedicated bioenergy crops in the UK." Renewable and Sustainable Energy Reviews 13(1): 271-290. RSB (2009). RSB Principles & Criteria for Sustainable Biofuel Production. Lausanne, Roundtable on Sustainable Biofuels: 28. RSB (2010). RSB Principles & Criteria for Sustainable Biofuel Production. Version 2.0. Lausanne, Roundtable on Sustainable Biofuels. SA Cane Growers association (2010). 2009/10 report of the board of directors of the South African cane growers' association. KwaZulu-Natal, South African Sugar Association. Sala, O. E., F. S. Chapin, III, J. J. Armesto, E. Berlow, J. Bloomfield, R. Dirzo, E. Huber-Sanwald, L. F. Huenneke, R. B. Jackson, A. Kinzig, R. Leemans, D. M. Lodge, H. A. Mooney, M. Oesterheld, iacute, N. L. Poff, M. T. Sykes, B. H. Walker, M. Walker and D. H. Wall (2000). "Global Biodiversity Scenarios for the Year 2100 ." Science 287(5459): 1770-1774. Sala, O. E., D. Sax and H. Leslie (2009). Biodiversity consequences of biofuel production. Biofuels: Environmental Consequences and Interactions with Changing Land Use. Proceedings of the Scientific Committee on Problems of the Environment (SCOPE) International Biofuels Project Rapid Assessment, 22-25 September 2008, Gummersbach Germany. R. W. Howarth and S. Bringezu. Ithaca NY, USA., Cornell University: 127-137. Salassi, M. E. and M. A. Deliberto (2007). Projected Commodity Costs And Returns - Sugarcane Production in Louisiana. Baton Rouge, LA, Farm Management Research & Extension Department of Agricultural Economics & Agribusiness. Salassi, M. E. and M. A. Deliberto (2008). Projected Commodity Costs And Returns - Sugarcane Production in Louisiana. Baton Rouge, LA, Farm Management Research & Extension Department of Agricultural Economics & Agribusiness. Salassi, M. E. and M. A. Deliberto (2009). Projected Commodity Costs And Returns - Sugarcane Production in Louisiana Baton Rouge, LA, Farm Management Research & Extension Department of Agricultural Economics & Agribusiness.

356

References

Salassi, M. E. and M. A. Deliberto (2010). Projected Commodity Costs And Returns - Sugarcane Production in Louisiana. . Baton Rouge, LA, Farm Management Research & Extension Department of Agricultural Economics & Agribusiness. Salassi, M. E. and M. A. Deliberto (2011). Projected Commodity Costs And Returns - Sugar cane production in Louisiana. Baton Rouge, LA, Farm management research & extension, Department of agricultural economics & agribusiness. Salassi, M. J. and J. Breux (2006). Projected Commodity Costs and Returns - Sugar cane production in Luisiana. Baton Rouge, LA, Louisiana State University Agricultural centre, department of agricultural economics and agribusines. Sanders, J. P. M. and D. van der Hoeven (2008). "Opportunities for a Bio-based Economy in the Netherlands." Energies 1(3): 105-119. Sang, T. and W. X. Zhu (2011). "China's bioenergy potential." Global Change Biology Bioenergy 3(2): 79-90. Savcor (2006). Investment in prospective of forest plantations in Mozambique. Helsinki, Finland, Savcor Indufor Oy. Schlegel, S., B. S. Elbersen and T. Kaphengst (2007). Input to DG TREN consultation on “Biofuel issues in the new legislation on the promotion of renewable energy” How to avoid major biodiversity loss from land use change induced by biomass production for bioenergy?, Ecologic – Institute for International and European Environmental Policy, Alterra. Schmit, C., M. D. A. Rounsevell and I. La Jeunesse (2006). "The limitations of spatial land use data in environmental analysis." Environmental Science & Policy 9(2): 174-188. Schneckenberger, K. and Y. Kuzyakov (2007). "Carbon sequestration under Miscanthus in sandy and loamy soils estimated by natural 13C abundance." Journal of Plant Nutrition and Soil Science 170(4): 538-542. Schouten, M., P. Verweij, A. Barendregt, R. Kleukers, V. Kalkman and P. de Ruiter (2009). "Determinants of species richness patterns in the Netherlands across multiple taxonomic groups." Biodiversity and Conservation 18(1): 203-217. Schreuder, R., W. v. Dijk, P. v. Asperen, J. d. Boer and J. R. v. d. Schoot (2008). Mebot 1.01 Beschrijving van Milieu- en bedrijfsmodel voor de Open Teelten. Lelystad, Nederland, Praktijkonderzoek Plant & Omgeving B.V.: 103. Seabra, J. E. A. and I. C. Macedo (2011). "Comparative analysis for power generation and ethanol production from sugarcane residual biomass in Brazil." Energy Policy 39(1): 421-428. Secretariat of the Convention on Biological Diversity and Netherlands Environmental Assessment Agency (2007). Cross-roads of Life on Earth — Exploring means to meet the 2010 Biodiversity Target. Solution oriented scenarios for Global Biodiversity Outlook 2. Montreal, Secretariat of the Convention on Biological Diversity: 90. Semere, T. and F. M. Slater (2007). "Invertebrate populations in miscanthus (Miscanthus×giganteus) and reed canary-grass (Phalaris arundinacea) fields." Biomass and Bioenergy 31(1): 30-39. Shell. (2008). "website; products and services, prices motorfuels." 2008, from http://www.shell.nl/home/content/nld/products_services/on_the_road/fuels/fuel_pricing/c pp/pricestructure/structure.html. . Shmida, A. and M. V. Wilson (1985). "Biological Determinants of Species Diversity." Journal of Biogeography 12(1): 1-20. Sikkema, R., M. Junginger, W. Pichler, S. Hayes and A. P. C. Faaij (2011a). "The international logistics of wood pellets for heating and power production in Europe: Costs, energy-input and greenhouse gas balances of pellet consumption in Italy, Sweden and the Netherlands (vol 5, pg 226, 2011)." Biofuels Bioproducts & Biorefining-Biofpr 5(2): 226-226. Sikkema, R., M. Steiner, M. Junginger, W. Hiegl, M. T. Hansen and A. Faaij (2011b). "The European wood pellet markets: current status and prospects for 2020." Biofuels, Bioproducts and Biorefining 5(3): 250-278. 357

SHADES OF GREEN

Smeets, E., M. Junginger, A. Faaij, A. Walter, P. Dolzan and W. Turkenburg (2008). "The sustainability of Brazilian ethanol--An assessment of the possibilities of certified production." Biomass and Bioenergy 32(8): 781-813. Smeets, E. M. W., L. F. Bouwman, E. Stehfest, Vuuren D. P. van and A. Posthuma (2009a). "Contribution of N2O to the greenhouse gas balance of first-generation biofuels." Global Change Biology 15(1): 1-23. Smeets, E. M. W. and A. P. C. Faaij (2010). "The impact of sustainability criteria on the costs and potentials of bioenergy production - Applied for case studies in Brazil and Ukraine." Biomass and Bioenergy 34(3): 319-333. Smeets, E. M. W., A. P. C. Faaij, I. M. Lewandowski and W. C. Turkenburg (2007). "A bottom-up assessment and review of global bio-energy potentials to 2050." Progress in Energy and Combustion Science 33(1): 56-106. Smeets, E. M. W., I. M. Lewandowski and A. P. C. Faaij (2009b). "The economical and environmental performance of miscanthus and switchgrass production and supply chains in a European setting." Renewable and Sustainable Energy Reviews 13(6-7): 1230-1245. Smith, P., D. Martino, Z. Cai, D. Gwary, H. Janzen, P. Kumar, B. McCarl, S. Ogle, F. O'Mara, C. Rice, B. Scholes, O. Sirotenko, M. Howden, T. McAllister, G. Pan, V. Romanenkov, U. Schneider, S. Towprayoon, M. Wattenbach and J. Smith (2008). "Greenhouse gas mitigation in agriculture." Philosophical Transactions of the Royal Society B: Biological Sciences 363(1492): 789-813. Smith, P., D. S. Powlson, J. U. Smith, P. Falloon and K. Coleman (2000). "Meeting Europe's climate change commitments: quantitative estimates of the potential for carbon mitigation by agriculture." Global Change Biology 6(5): 525-539. State statistic service Ukraine (2011a). Demographic statistics by region. State statistic service Ukraine (2011b). ILO unemployment rate of population in 2011, by region (information is based on results of sapling survey population of economic activity). State Statistics Service of Ukraine (2011). Agriculture in Ukriane, Statistical Yearbook 2010. Kiev. Statline. (2008a). "Aardgas en elektriciteit; gemiddelde tarieven." from http://statline.cbs.nl/StatWeb/selection/?DM=SLNL&PA=37359&VW=T. Statline. (2008b). "Heffingen op energiedragers." 2008. Statline. (2008c). "Huisbrandolie, petroleum en zware stookolie; gemiddelde adviesprijzen ", from http://statline.cbs.nl/StatWeb/selection/?DM=SLNL&PA=7520&VW=T. Statline. (2008d). "Motorbrandstoffen; adviesprijzen." Stehfest, E. and L. Bouwman (2006). "N2O and NO emission from agricultural fields and soils under natural vegetation: summarizing available measurement data and modeling of global annual emissions." Nutrient Cycling in Agroecosystems 74(3): 207-228. Strengers, B., R. Leemans, B. Eickhout, B. de Vries and L. Bouwman (2004). "The land-use projections and resulting emissions in the IPCC SRES scenarios scenarios as simulated by the IMAGE 2.2 model." GeoJournal 61(4): 381-393. Styles, D. and M. B. Jones (2007a). "Current and future financial competitiveness of electricity and heat from energy crops: A case study from Ireland." Energy Policy 35(8): 4355-4367. Styles, D. and M. B. Jones (2007b). "Energy crops in Ireland: Quantifying the potential life-cycle greenhouse gas reductions of energy-crop electricity." Biomass and Bioenergy 31(11-12): 759-772. Styles, D., F. Thorne and M. B. Jones "Energy crops in Ireland: An economic comparison of willow and Miscanthus production with conventional farming systems." Biomass and Bioenergy 52(5): 407-421. Tasser, E., J. Walde, U. Tappeiner, A. Teutsch and W. Noggler (2007). "Land-use changes and natural reforestation in the Eastern Central Alps." Agriculture, Ecosystems & Environment 118(1-4): 115-129. 358

References

Timberlake, J. R. and S. J. Reddy (1986). Potential livestock productivity and livestock carrying capacity over Mozambique Cominicação No. 49. Maputo, Mozambique, Do Instituto Nacional de investigação Agrinomica. Tipper, R., C. Hutchison and M. Brander (2009). A Practical Approach for Policies to Address GHG Emissions from Indirect Land Use Change Associated with Biofuels. Edinburgh, Ecometrica. Tostao, E. and B. W. Brorsen (2005). "Spatial Efficiency in Mozambique’s Post-Reform Maize Markets." Agricultural Economics 33(02): 205-214. Uaiene, R. (2004). Maize and Sorghum technologies and the effects of marketing strategies on farmers' income in Mozambique. West Lafayette, Indianapolis, USA, Purdue University. Master of Science: 139. Ugalde, L., O. Pérez and D. J. Mead (Editor) (2001). Mean Annual Volume Increment of Selected Industrial Forest Plantation Species. Rome, Forestry Department, Food and Agriculture Organization of the United Nations. UNDP (2004). Water Governance for Poverty Reduction Key Issues and the UNDP Response to Millenium Development Goals. New York, United Nations Development Programme: 93. UNDP (2008). World Population Prospects: The 2008 Revision, Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat. UNDP (2009). World Population Prospects: The 2008 Revision. New York, United Nations, Population Division of the Department of Economic and Social Affairs of the United Nations Secretariat. UNDP. (2011, 2 January 2012). "World Population Prospects: The 2010 Revision." from http://esa.un.org/unpd/wpp/index.htm UNDP, UNDESA and WEC (2000). World energy Assessment. New York, United Nations Develpoment Programm, United Nations Department of Economic and Social Affairs, World Energy Outlook. UNDP, UNDESA and WEC (2004). World Energy Assessment overview 2004 Update. J. G. a. T. B. Johansson. London, United Nations Development Programme, United Nations Department of Economic and Social Affairs, World Energy Council, . UNEP (2002). Global Environment Outlook 3. Nairobi, Kenia, United Nations Environment Programme. UNFCC (2011). National Inventory Submissions 2011: Annex I Party GHG Inventory Submissions. Common Reporting Format: Ukraine, United Nations Framework Convention on Climate Change. USAID (2006). Environmental and economic impact assessment of land privatisation in eastern Europe. Burlington, National Agrarian University of Ukraine - Institute of LandResources, MPC Consulting, ARD, Inc. USAID (2008). Maputo Corridor Summary Report A Transport Logistics Diagnostic Tool Study March. Washington, USA, USAID. USDA and NRCS. (1996). "Soil Quality Resource Concerns: Soil Erosion." from http://soils.usda.gov/sqi/publications/files/sq_two_1.pdf. USDA and NRCS (2002). National Agronomy Manual Washington, United States Department of Agriculture, Natural Resource Conservation Service. Uslu, A., A. P. C. Faaij and P. C. A. Bergman (2008). "Pre-treatment technologies, and their effect on international bioenergy supply chain logistics. Techno-economic evaluation of torrefaction, fast pyrolysis and pelletisation." Energy 33(8): 1206-1223. van Dam, J., A. P. C. Faaij, L. Gustavsson and W. C. Turkenburg (2010a). A unified methodology and tool to evaluate greenhouse gas balances and cost-effectiveness of various biomass energy systems. Chapter 6. Sustainability of Bioenergy chains: the result is in the details. PhD Thesis. J. v. Dam. Utrecht, Copernicus Institute, Utrecht University. van Dam, J., A. P. C. Faaij, J. Hilbert, H. Petruzzi and W. C. Turkenburg (2009a). "Large-scale bioenergy production from soybeans and switchgrass in Argentina: Part A: Potential and 359

SHADES OF GREEN

economic feasibility for national and international markets." Renewable and Sustainable Energy Reviews 13(8): 1710-1733. van Dam, J., A. P. C. Faaij, J. Hilbert, H. Petruzzi and W. C. Turkenburg (2009b). "Large-scale bioenergy production from soybeans and switchgrass in Argentina: Part B. Environmental and socio-economic impacts on a regional level." Renewable and Sustainable Energy Reviews 13(8): 1679-1709. van Dam, J., A. P. C. Faaij, I. Lewandowski and G. Fischer (2007). "Biomass production potentials in Central and Eastern Europe under different scenarios." Biomass and Bioenergy 31(6): 345366. van Dam, J., M. Junginger and A. P. C. Faaij (2010b). "From the global efforts on certification of bioenergy towards an integrated approach based on sustainable land use planning." Renewable and Sustainable Energy Reviews 14(9): 2445-2472. van Dam, J., M. Junginger, A. P. C. Faaij, I. Jürgens, G. Best and U. Fritsche (2008). "Overview of recent developments in sustainable biomass certification." Biomass and Bioenergy 32(8): 749-780. van den Broek, R., S. Teeuwisse, K. Healion, T. Kent, A. van Wijk, A. Faaij and W. Turkenburg (2001). "Potentials for electricity production from wood in Ireland." Energy 26(11): 991-1013. van den Broek, R., T. van den Burg, A. van Wijk and W. Turkenburg (2000a). "Electricity generation from eucalyptus and bagasse by sugar mills in Nicaragua: A comparison with fuel oil electricity generation on the basis of costs, macro-economic impacts and environmental emissions." Biomass and Bioenergy 19(5): 311-335. van den Broek, R., A. van Wijk and W. Turkenburg (2000b). "Farm-based versus industrial eucalyptus plantations for electricity generation in Nicaragua." Biomass and Bioenergy 19(5): 295-310. van den Broek, R., A. van Wijk and W. Turkenburg (2002). "Electricity from energy crops in different settings--a country comparison between Nicaragua, Ireland and the Netherlands." Biomass and Bioenergy 22(2): 79-98. van den Wall Bake, J. D., M. Junginger, A. Faaij, T. Poot and A. Walter (2009). "Explaining the experience curve: Cost reductions of Brazilian ethanol from sugarcane." Biomass and Bioenergy 33(4): 644-658. Van der Hilst, F., V. Dornburg, J. P. M. Sanders, B. Elbersen, A. Graves, W. C. Turkenburg, H. W. Elbersen, J. M. C. Van Dam and A. P. C. Faaij (2010). "Potential, spatial distribution and economic performance of regional biomass chains: The North of the Netherlands as example." Agricultural Systems 103(7): 403-417. Van der Hilst, F., J. M. C. Van Dam, P. A. Verweij, M. J. P. M. Riksen, J. P. M. Sanders and A. P. C. Faaij (2012a). "Spatial variation in environmental impacts of bioenergy supply chains." RSER 16(2012): 2053-2069. Van der Hilst, F., J. A. Verstegen and A. P. C. Faaij (2012b). Dynamic and integated modeling of GHG emision from the implementation of energy crop production and intensification of the agricultural sector. Utrecht, Copernicus Institute, Utrecht University. Van der Hilst, F., J. A. Verstegen, D. Karssenberg and A. P. C. Faaij (2011). "Land use change modeling for the assessment of land availability for energy crops: the case of Mozambique " Global Change Biology Bionenergy In press DOI: 10.1111/j.1757-1707.2011.01147.x. van der Voort, M. P. J., R. D. Timmer, W. v. Geel, W. Runia and W. J. Corre (2008). Economie van Energiegewassen. Wageningen, The Netherlands, Praktijkonderzoek plant & omgeving. Akkerbouw, groene ruimte en vollegrondsgroente, Wageningen Universiteit en research centrum: 79. van Eijck, J., E. M. W. Smeets and A. P. C. Faaij (2012). "The economic performance of Jatropha, Cassava and Eucalyptus production systems for energy in an East African smallholder setting." Global Change Biology Bioenergy Accepted.

360

References

van Kerckhoven, S., M. J. P. M. Riksen and W. Cornelis (2009). Afbakening van gebieden gevoelig aan winderosie in Vlaanderen. Eindrapport. Gent, Belgium, Vakgroep Bodembeheer - Faculteit Bio-enginieurswetenschappen. Gent University: 79. van Rooij, W. (2008). Manual for biodiversity modelling on a national scale. Using GLOBIO3 and CLUE methodology to calculate current and future status of biodiversity. Bilthoven, The Netherlands, PBL. Veldkamp, A. and L. O. Fresco (1996). "CLUE: a conceptual model to study the Conversion of Land Use and its Effects." Ecological Modelling 85(2-3): 253-270. Veldkamp, A. and E. F. Lambin (2001). "Predicting land-use change." Agriculture, Ecosystems & Environment 85(1-3): 1-6. Veldkamp, A. and P. H. Verburg (2004). "Modelling land use change and environmental impact." Journal of Environmental Management 72(1-2): 1-3. Velthof, G. L. and P. J. Kuikman (2000). Beperking van lachgasemissie uit gewasresten. Een systeemanalyse. Wageningen, Alterra: 81. Velthof, G. L., D. Oudendag and O. Oenema (2007). Development and application of the integrated nitrogen model MITERRA-EUROPE. Wageningen, Alterra: 102. Velthof, G. L., D. Oudendag, H. P. Witzke, W. A. H. Asman, Z. Klimont and O. Oenema (2009). "Integrated Assessment of Nitrogen Losses from Agriculture in EU-27 using MITERRAEUROPE." J Environ Qual 38(2): 402-417. Venturi, P., J. K. Gigler and W. Huisman (1999). "Economical and technical comparison between herbaceous (Miscanthus x giganteus) and woody energy crops (Salix viminalis)." Renewable Energy 16(1-4): 1023-1026. Venturi, P., W. Huisman and J. Molenaar (1996). Cost calculations of production chains of Miscanthus Giganteus. Wageningen, Department of Agricultural Engineering and Physics, Wageningen Agricultural University: 45. Verburg, P. and K. Overmars (2009). "Combining top-down and bottom-up dynamics in land use modeling: exploring the future of abandoned farmlands in Europe with the Dyna-CLUE model." Landscape Ecology 24(9): 1167-1181. Verburg, P. H., G. H. J. de Koning, K. Kok, A. Veldkamp and J. Bouma (1999). "A spatial explicit allocation procedure for modelling the pattern of land use change based upon actual land use." Ecological Modelling 116(1): 45-61. Verburg, P. H., B. Eickhout and H. van Meijl (2008). "A multi-scale, multi-model approach for analyzing the future dynamics of European land use." Annals of Regional Science 42(1): 5777. Verburg, P. H., P. P. Schot, M. J. Dijst and A. Veldkamp (2004). "Land use change modelling: current practice and research priorities." GeoJournal 61(4): 309-324. Verburg, P. H., C. J. E. Schulp, N. Witte and A. Veldkamp (2006). "Downscaling of land use change scenarios to assess the dynamics of European landscapes." Agriculture Ecosystems & Environment 114(1): 39-56. Verburg, R., E. Stehfest, G. Woltjer and B. Eickhout (2009). "The effect of agricultural trade liberalisation on land-use related greenhouse gas emissions." Global Environmental ChangeHuman and Policy Dimensions 19(4): 434-446. Verstegen, J. A., D. Karssenberg, F. van der Hilst and A. Faaij (2011). "Spatio-temporal uncertainty in Spatial Decision Support Systems: A case study of changing land availability for bioenergy crops in Mozambique." Computers, Environment and Urban Systems I36(1): 30-42. Walsh, M. E., D. G. D. Ugarte, H. Shapouri and S. P. Slinsky (2003). "Bioenergy crop production in the United States - Potential quantities, land use changes, and economic impacts on the agricultural sector." Environmental & Resource Economics 24(4): 313-333.

361

SHADES OF GREEN

Walsh, S. J., J. P. Messina, C. F. Mena, G. P. Malanson and P. H. Page (2008). "Complexity theory, spatial simulation models, and land use dynamics in the Northern Ecuadorian Amazon." Geoforum 39(2): 867-878. Walter, A., P. Dolzan and E. Piacente (2006). Biomass Energy and Bio-energy Trade: Historic Developments in Brazil and Current Opportunities. Country Report: Brazil – Task 40 – Sustainable Bio-energy Trade; securing Supply and Demand Final Version. State University of Campinas – Unicamp. Campinas – Brazil. Wassenaar, T., P. Gerber, P. H. Verburg, M. Rosales, M. Ibrahim and H. Steinfeld (2007). "Projecting land use changes in the Neotropics: The geography of pasture expansion into forest." Global Environmental Change 17(1): 86-104. Watson, H. K. (2011). "Potential to expand sustainable bioenergy from sugarcane in southern Africa." Energy Policy 39(10): 5746-5750. Watson, H. K., G. G. Garland, B. Purchase, N. Dercas, P. Griffee and F. X. Johnson (2008). Bioenergy for Sustainable Development and Global Competitiveness: the case of Sugar Cane in Southern Africa A compilation of Results from the Thematic Research Network:Cane Resources Network for Southern Africa (CARENSA). Thematic Report 1: Agriculture. Stockholm, Sweden, Stockholm Environment Institute: 59. West, T. O. and G. Marland (2002). "A synthesis of carbon sequestration, carbon emissions, and net carbon flux in agriculture: comparing tillage practices in the United States." Agriculture, Ecosystems & Environment 91(1-3): 217-232. Westhoek, H. J., M. van den Berg and J. A. Bakkes (2006). "Scenario development to explore the future of Europe's rural areas." Agriculture, Ecosystems & Environment 114(1): 7-20. Wicke, B., E. Smeets, A. Tabeau, J. Hilbert and A. Faaij (2009). "Macroeconomic impacts of bioenergy production on surplus agricultural land--A case study of Argentina." Renewable and Sustainable Energy Reviews 13(9): 2463-2473. Wicke, B., D. P. v. Vuuren, P. A. Verweij, H. Van Meijl and A. P. C. Faaij (2012). "Indirect land use change: Review of existing models and strategies for mitigation." Biofuels 3(1): 87-100. World Bank (2006). Mozambique Agricultural Development Strategy Stimulating Smallholder Agricultural Growth. Washington DC, The world Bank, AFTS1 Agriculture, Environment, and Social Development Unit Country Department 2 Africa Region. World Bank and OECD (2004). Achieving Ukraine's Agricultural Potential Stimulating Agricultural Growth And Improving Rural Life. Washington, USA, World Bank, OECD. WSRG (1994). Harvesting Decision Support System. Aberdeen, Wood Supply Research group, Department of Forestry, University of Aberdeen. WSRG (1995). Short Rotation Forestry Handbook. D. o. F. Wood Supply Research group, University of Aberdeen. Aberdeen. WWI (2006). Biofuels for transportation. Global potential and implications for sustainable agriculture and energy in the 21st century. . Washington DC, USA, World Watch Institute. Xavier, C. E. O., L. B. Zilio, D. Y. Sonoda and P. V. Marques (2009). Custos de produçao de cana-deaçúcar, açúcar e etanol no Brasil: 2008/2009. São Paulo, Universidade de São Paulo - USP, Escola Superior de Agricultura “Luiz de Queiroz” - ESALQ, Departamento de Economia, Administração e Sociologia, Programa de Educação Continuada em Economia e Gestão de empresas - PECEGE. You, L., S.Crespo, Z. Guo, J. Koo, W. Ojo, K. Sebastian, M.T. Tenorio, S. Wood, U. Wood-Sichra. S (2006). patial Produciton Allocation Model (SPAM) 2000 Version 3 Release 2. Zavale, H., E. Mabaya and R. Christy (2006). Smallholders’ Cost Efficiency in Mozambique: Implications for Improved Maize Seed Adoption. Contributed Paper Prepared for Presentation at the International Association of Agricultural Economists Conference, Gold Coast, Australia,.

362

References

Zuidema, G., G. J. Born, J. Alcamo and G. J. J. Kreileman (1994). "Simulating changes in global land cover as affected by economic and climatic factors." Water, Air, & Soil Pollution 76(1): 163-198. Zuurbier, P. and J. van der Vooren (2008). Sugarcane ethanol: Contributions to climate change mitigation and the environment. Wageningen, Wageningen Academic Publishers.

363

Dankwoord

Dankwoord De weg naar een PhD is zelden een rechte lijn en ook mijn route kende wat detours, potholes, pieken en dalen. Ik ben erg blij dat het boekje nu af is maar dat betekent niet dat ik niet genoten heb van de trip. Het zijn juist de mensen die mij onderweg geholpen hebben die de reis de moeite waard hebben gemaakt. Allereerst André, dank voor je enorme vertouwen en de vrijheid die je me hebt gegund. Je gaf me de ruimte mijn ideeën uit te werken, op pad te gaan, mijn plan te trekken. Daarnaast gaf je me ook de steun op de momenten dat het nodig was: je wees me de rode draad waar ik verdwaald raakte in de details, je haalde je schouders op voor de beren die ik op de weg zag en je dacht altijd in oplossingen. Je enthousiasme is aanstekelijk en werkt buitengewoon motiverend. Johan, ik heb onze discussies erg gewaardeerd en deze hebben me geleerd mijn keuzes goed te verantwoorden. Je grote vermogen om ‘out of the box’ te denken betekent een aanwinst voor de biobased economy. Ik heb genoten van onze gesprekken die van onderzoek, via de huidige politieke ontwikkelingen en fietsen in Frankrijk, weer terug kwamen bij waar het nou eigenlijk heen moet met de wereld. Wim, drie bleek teveel volgens het promotiereglement maar ik ben je erg dankbaar voor je begeleiding en adviezen aan het begin van mijn promotie en je bijdrage aan het eerste artikel. Veronica en Jinke, bedankt voor jullie begeleiding in het eerste stadium en jullie bijdrage aan de eerste twee artikelen. Juist in deze fase was het erg fijn om sparringpartners te hebben. Jacco, dank voor het klankboorden in het eerste jaar. Dank aan al mijn collega’s van het ME4 projectteam: Bert, Wolter, Hans, Ria, Hamid, Marc, Paul, Eefje, Matthijs, Jan-Peter, Berien, Igor, Peter en Janneke. Veel dank gaat uit naar de collega’s voor hun bijdrage aan de eerste twee artikelen: Jan-Peter, Berien, Wolter, Michiel en Rob. Jan-Peter, dank dat je altijd klaar stond voor vragen over Miterra. Marc, heel fijn dat jij mijn maatje was binnen dit project. Judith, dankzij jou hebben we niet alleen de ruimtelijk expliciete, maar ook de dynamische dimensie aan het onderzoek kunnen geven. Je scherpte, je enorme werklust en je humor maakte het erg prettig en efficiënt samenwerken. Ik ben erg blij dat je nu in het Be-Basic project zit en ik kijk uit naar de gezamenlijke projecten. Derek, toen ik bij jou op de deur kwam kloppen voor eventuele samenwerking was je direct enthousiast en hebben we er

365

SHADES OF GREEN

snel werk van kunnen maken. Leuk dat het Mozambique project een voorzet is gebleken voor verdere samenwerking. Maarten, dank voor al je hulp met GIS. Ik wil al mijn huidige en oude Utrechtse collega’s bedanken voor de fijne werkomgeving. Dank aan Birka, Marc, Martin, Janske, Ric, Bothwell en Edward voor de inhoudelijke discussies. Ric, Hans en Aisha dank voor jullie hulp bij alle hard- en software problemen. Birka, heel fijn dat je me hielp in de laatste fase van mijn proefschrift. Je scherpe blik was erg welkom. Janske, ik ben erg blij dat we kamergenoten zijn en dank voor al je support in de laatste maanden. Onze field trip in Mozambique is onvergetelijk en getuigde van echte team effort (hoe krijg je een 4x4 over een ingestorte brug?). Het lijkt me geweldig om met dit gouden team nog vaker op onderzoek uit te gaan! Natuurlijk ook dank aan de rest van de bio-energie groep en de collega’s van de biodiversiteit, CCS, PV, Risico en Demand groepen. Het is een voorrecht te werken met zo’n divers gezelschap van zeer toegewijde mensen. Aisha, Siham en Petra, dank voor al jullie hulp vooral in de laatste maanden en voor jullie gezelligheid. Gelukkig is er meer dan werk. De deelnames aan de Pheidippidesloop met het ‘Rainbow’ en ‘Sprinting To Science’ team, het schaatsen op strenge winterdagen en het fietsen in de zomer waren welkome afwisselingen. Dank aan mijn Wageningse collega’s. Het is niet altijd eenvoudig om simultaan te integreren in twee groepen van verschillende universiteiten. Ondanks dat mijn werk inhoudelijk wat verder van de groep aflag, heb ik me altijd erg welkom gevoeld. De sportdag, de teambuilding uitjes, de kerstdiners, en de speciale biertjes in de Vlaamse reus waren erg gezellig. Gerda, dank voor al je hulp aan de Wageningse kant. My visits to Mozambique were truly inspiring. I’m so grateful for all the people I met there and everyone who helped me out with finding my way and contributed to gathering knowledge, information and data. Many thanks to the people of DNTF, DNER, CEPAGRI and IIAM for sharing their knowledge and data and to all the biofuel, forestry and agricultural projects we visited for sharing their experiences. Olivio and Rita, Cristóvão, and Mario, thank you for all the diners and drinks. You made me feel part of the ‘family’. Marc dank voor je hulp en gezelligheid in Mozambique! Leuk dat we nu bijna op hetzelfde moment promoveren. Ik kijk uit naar het triathlonseizoen! Vasco, Leticia, Pedro en Camilla, thank you for your contributions to the Portuguese translation of the summary and conclusions. Thanks for finding time in you busy schedules. Thanks to the people of SEC biomass for hosting me in Kiev. Tetiana and Olga, I’m very grateful for your contribution to the paper on bioenergy in Ukraine (Chapter 6) and for 366

Dankwoord

translating the summary and conclusions of this thesis in Ukrainian. Olga, Ina and Tanya, I really enjoyed the tours and our dinners and drinks in Kiev. Thanks for your hospitality. Dank aan al mijn fiets-, zwem-, hardloop- en alpine- maatjes die ervoor hebben gezorgd dat ik regelmatig mijn hoognodige dosis endorfines binnenkreeg in combinatie met een leeg hoofd en een hoop gezelligheid. Dit maakt dat ik weer met frisse ideeën, ontspannen lijf en hernieuwde energie achter mijn laptop kroop. V123 +, dank voor alle onzin die het leven zin geeft! D’01, fijn dat we ook nu we niet meer in hetzelfde schuitje zitten, hoogte- en dieptepunten met elkaar te delen. El, leuk om met jou de passie voor sport, onderzoek en reizen te delen. We-devils, wat goed dat we elkaar blijven zien. Karen, dank voor je kritische blik op het Oekraïne artikel, gelukkig is Londen niet zo ver. Erika en Anouk, jullie hebben de ups en downs van mijn PhD traject van dichtbij meegekregen. Ik vind het erg leuk dat jullie mij ook in die laatste 45 minuten bijstaan en mijn (para)nymf(oman)en zijn:). Rink, het volgende boek schrijven we samen. Met ook wat grafiekjes, een paar zwarte bladzijden en een wat boeiendere verhaallijn. Lieve Moen, je bent er altijd, voor alles! Daar ben ik je heel dankbaar voor. Lieve pap, mam, Sanne, Roos en Benjamin, dank voor jullie steun. Sanne en Ivo fijn dat jullie altijd in zijn voor een drankje/etentje/spelletje. Ik verheug me op het nieuwe hoofdstuk dat nu aanbreekt! Lieve Roos, grappig te merken dat de PhD issues, vakgebied-, universiteit- en landoverstijgend zijn. Ik mis je erg nu in je in Oslo zit. Magnus and Roos, looking forward to all the coming visits! Bennie, je hebt dan dezelfde richting gekozen, je pakt het op geheel eigen wijze aan. Mooi hoe je samen met Esmee studie, werk, ontspanning en avonturen weet te combineren. Lieve pap en mam, dank voor jullie nooit aflatende trots en voor alle vrijheid die jullie mij van jongs af aan hebben gegeven om te gaan en staan waar ik wil en mijn eigen keuzes te maken. Ik had me geen betere voorbereiding op het promoveren kunnen wensen dan: ‘leer mij het zelf te doen’. Lieve Joost, met jou is alles leuker! Ik kijk uit naar al onze volgende avonturen.

367

Curriculum Vitae

Curriculum Vitae Floor van der Hilst was born on September 28, 1981 in Utrecht, the Netherlands. She studied Science and Innovation management at Utrecht University specializing in Energy and Resources and taking additional courses in Life Sciences and in Sustainable Development. During her Bachelor and Masters, she received an athletic scholarship for her achievements in rowing and she was a member of several students’ boards and committees. Her MSc thesis focused on the implementation of micro combined heat and power systems. She finished her Masters with honours in the beginning of 2005 after which she worked as a teaching assistant and travelled in Latin America. End of 2005, she started as junior researcher in the department of innovation and environmental science at Utrecht University, and contributed to a European project on sustainable transportation and several MSc courses. From 2006 to 2007 she worked as scientific employee in the Policy Studies Unit of the Energy research Centre of the Netherlands. In 2007 Floor started her PhD research which was set up as a joint project between the Science, Technology & Society (STS) group of the Copernicus Institute of Utrecht University and the Valorisation of Plant Production chains group of Wageningen University. The PhD project was part of the ME4 project of the Climate changes spatial planning program. Her research focused on the sustainability of bioenergy production, with special attention for the avoidance of indirect land use changes, the environmental impacts and economic viability. Her work combines methodology and model development with field work experience. For her PhD research she spent several working periods in Mozambique and a brief one in Ukraine. She presented her work at several scientific conferences and meetings in the Netherlands, Belgium, Germany, Spain, Ukraine, Mozambique and Brazil. During her PhD she contributed to several BSc and MSc courses related to energy and sustainable development and supervised MSc thesis projects. Floor recently started as a postdoctoral researcher at the Copernicus Institute.

369

Loading...

shades of green - Utrecht University Repository - Universiteit Utrecht

SHADES OF GREEN Spatial and temporal variability of potentials, costs and environmental impacts of bioenergy production SHADES OF GREEN Spatial and ...

33MB Sizes 1 Downloads 0 Views

Recommend Documents

televising locality - Utrecht University Repository - Universiteit Utrecht
http://www.kpi.go.id/download/laporan_tahunan/LAT_KPI_2011_Final.pdf (accessed on 12 June 2012). 7 I received the data f

COVER OKE.cdr - Utrecht University Repository - Universiteit Utrecht
Giving in Islam and Approaches to the Study of Charities. 28. Islamic Forms of Giving .... 232 Figure 11 One ustadh teac

Neglected Voices - Utrecht University Repository - Universiteit Utrecht
Aug 13, 2013 - verschuivingen kolonialisme, kerstening, gewapend conflict tijdens de Mulele rebellie en dertig jaar poli

4na'R2 - Utrecht University Repository - Universiteit Utrecht
synopsis. A simple analytical expression for the absolute ionization and (optically allowed) excitation cross sections,

Di ki manera? - Utrecht University Repository - Universiteit Utrecht
Feb 13, 2007 - Dr. Alan. Benjamin provided much support and encouragement. He read the manuscript of the the- sis with a

Quaestiones Infinitiae - Utrecht University Repository - Universiteit
37 Among the required books figure Porphyry's Isagoge (3rd century), which was an introduction to Aristotle's Categoriae

ERPL_14-1 knipfile.qxd - Utrecht University Repository - Universiteit
Zusammenfassung: Ursprünglich sah man Privatrecht als immun vom Einfluß der. Grundrechte an, deren Funktion darauf bes

dr. Rak Kim - Geosciences - Utrecht University - Universiteit Utrecht
From 2014 to 2017, Kim was Book Review Editor of Transnational Environmental Law (Cambridge University Press). In additi

a thesis - Universiteit Utrecht
jawaban kelompoknya masing-masing dan melakukan koreksi sehingga terbentuk kesadaran pada siswa mana jawaban yang tepat.

Introduction P - Universiteit Utrecht
This thematic issue of P: Portuguese Cultural Studies focuses on the interactions between critiques of colonialism and c